Updating ProkBert class and adding LCA tokenizer config

config.json

@@ -1,7 +1,6 @@
 {
-  "_name_or_path": "/scratch/fastscratch/NBL/trained_models/test/prokbert-mini-k6s1/checkpoint-470000",
   "architectures": [
-    "MegatronBertForMaskedLM"
+    "ProkBertForMaskedLM"
   ],
   "attention_probs_dropout_prob": 0.1,
   "hidden_act": "gelu",
@@ -9,15 +8,17 @@
   "hidden_size": 384,
   "initializer_range": 0.02,
   "intermediate_size": 3072,
+  "kmer": 6,
   "layer_norm_eps": 1e-12,
   "max_position_embeddings": 1024,
-  "model_type": "megatron-bert",
+  "model_type": "prokbert",
   "num_attention_heads": 6,
   "num_hidden_layers": 6,
   "pad_token_id": 0,
   "position_embedding_type": "relative_key_query",
+  "shift": 1,
   "torch_dtype": "float32",
-  "transformers_version": "4.30.2",
+  "transformers_version": "4.33.2",
   "type_vocab_size": 2,
   "use_cache": true,
   "vocab_size": 4101

config_utils.py

@@ -1,443 +0,0 @@
-# Config utils
-import yaml
-import pathlib
-from os.path import join
-import os
-import numpy as np
-import torch
-from multiprocessing import cpu_count
-
-class BaseConfig:
-    """Base class for managing and validating configurations."""
-
-    numpy_dtype_mapping = {1: np.int8,
-                           2: np.int16,
-                           8: np.int64,
-                           4: np.int32}
-
-    def __init__(self):
-        super().__init__()
-
-    def cast_to_expected_type(self, parameter_class: str, parameter_name: str, value: any) -> any:
-        """
-        Cast the given value to the expected type.
-
-        :param parameter_class: The class/category of the parameter.
-        :type parameter_class: str
-        :param parameter_name: The name of the parameter.
-        :type parameter_name: str
-        :param value: The value to be casted.
-        :type value: any
-        :return: Value casted to the expected type.
-        :rtype: any
-        :raises ValueError: If casting fails.
-        """
-        expected_type = self.parameters[parameter_class][parameter_name]['type']
-
-        if expected_type in ["integer", "int"]:
-            try:
-                return int(value)
-            except ValueError:
-                raise ValueError(f"Failed to cast value '{value}' to integer for parameter '{parameter_name}' in class '{parameter_class}'.")
-        elif expected_type == "float":
-            try:
-                return float(value)
-            except ValueError:
-                raise ValueError(f"Failed to cast value '{value}' to float for parameter '{parameter_name}' in class '{parameter_class}'.")
-        elif expected_type in ["string", "str"]:
-            return str(value)
-        elif expected_type in ["boolean", "bool"]:
-            if isinstance(value, bool):
-                return value
-            elif str(value).lower() == "true":
-                return True
-            elif str(value).lower() == "false":
-                return False
-            else:
-                raise ValueError(f"Failed to cast value '{value}' to boolean for parameter '{parameter_name}' in class '{parameter_class}'.")
-        elif expected_type == "type":
-            # For this type, we will simply return the value without casting. 
-            # It assumes the configuration provides valid Python types.
-            return value
-        elif expected_type == "list":
-            if isinstance(value, list):
-                return value
-            else:
-                raise ValueError(f"Failed to validate value '{value}' as a list for parameter '{parameter_name}' in class '{parameter_class}'.")
-        elif expected_type == "tuple":
-            if isinstance(value, tuple):
-                return value
-            else:
-                raise ValueError(f"Failed to validate value '{value}' as a tuple for parameter '{parameter_name}' in class '{parameter_class}'.")
-        elif expected_type == "set":
-            if isinstance(value, set):
-                return value
-            else:
-                raise ValueError(f"Failed to validate value '{value}' as a set for parameter '{parameter_name}' in class '{parameter_class}'.")
-        elif expected_type == "dict":
-            if isinstance(value, dict):
-                return value
-            else:
-                raise ValueError(f"Failed to validate value '{value}' as a dict for parameter '{parameter_name}' in class '{parameter_class}'.")
-        else:
-            raise ValueError(f"Unknown expected type '{expected_type}' for parameter '{parameter_name}' in class '{parameter_class}'.")
-
-
-
-    def get_parameter(self, parameter_class: str, parameter_name: str) -> any:
-        """
-        Retrieve the default value of a specified parameter.
-
-        :param parameter_class: The class/category of the parameter (e.g., 'segmentation').
-        :type parameter_class: str
-        :param parameter_name: The name of the parameter.
-        :type parameter_name: str
-        :return: Default value of the parameter, casted to the expected type.
-        :rtype: any
-        """
-        default_value = self.parameters[parameter_class][parameter_name]['default']
-        return self.cast_to_expected_type(parameter_class, parameter_name, default_value)
-    
-
-    
-    def validate_type(self, parameter_class: str, parameter_name: str, value: any) -> bool:
-        """
-        Validate the type of a given value against the expected type.
-
-        :param parameter_class: The class/category of the parameter.
-        :type parameter_class: str
-        :param parameter_name: The name of the parameter.
-        :type parameter_name: str
-        :param value: The value to be validated.
-        :type value: any
-        :return: True if the value is of the expected type, otherwise False.
-        :rtype: bool
-        """
-        expected_type = self.parameters[parameter_class][parameter_name]['type']
-
-        if expected_type == "integer" and not isinstance(value, int):
-            return False
-        elif expected_type == "float" and not isinstance(value, float):
-            return False
-        elif expected_type == "string" and not isinstance(value, str):
-            return False
-        else:
-            return True
-    
-    def validate_value(self, parameter_class: str, parameter_name: str, value: any) -> bool:
-        """
-        Validate the value of a parameter against its constraints.
-
-        :param parameter_class: The class/category of the parameter.
-        :type parameter_class: str
-        :param parameter_name: The name of the parameter.
-        :type parameter_name: str
-        :param value: The value to be validated.
-        :type value: any
-        :return: True if the value meets the constraints, otherwise False.
-        :rtype: bool
-        """
-        constraints = self.parameters[parameter_class][parameter_name].get('constraints', {})
-        
-        if 'options' in constraints and value not in constraints['options']:
-            return False
-        if 'min' in constraints and value < constraints['min']:
-            return False
-        if 'max' in constraints and value > constraints['max']:
-            return False
-        return True
-    
-
-    def validate(self, parameter_class: str, parameter_name: str, value: any):
-        """
-        Validate both the type and value of a parameter.
-
-        :param parameter_class: The class/category of the parameter.
-        :type parameter_class: str
-        :param parameter_name: The name of the parameter.
-        :type parameter_name: str
-        :param value: The value to be validated.
-        :type value: any
-        :raises TypeError: If the value is not of the expected type.
-        :raises ValueError: If the value does not meet the parameter's constraints.
-        """
-        if not self.validate_type(parameter_class, parameter_name, value):
-            raise TypeError(f"Invalid type for {parameter_name} for parameter class '{parameter_class}'. Expected {self.parameters[parameter_class][parameter_name]['type']}.")
-        
-        if not self.validate_value(parameter_class, parameter_name, value):
-            raise ValueError(f"Invalid value for {parameter_name}  for parameter class '{parameter_class}'. Constraints: {self.parameters[parameter_class][parameter_name].get('constraints', {})}.")
-
-    def describe(self, parameter_class: str, parameter_name: str) -> str:
-        """
-        Retrieve the description of a parameter.
-
-        :param parameter_class: The class/category of the parameter.
-        :type parameter_class: str
-        :param parameter_name: The name of the parameter.
-        :type parameter_name: str
-        :return: Description of the parameter.
-        :rtype: str
-        """
-        return self.parameters[parameter_class][parameter_name]['description']
-
-
-
-class SeqConfig(BaseConfig):
-    """Class to manage and validate sequence processing configurations."""
-
-    def __init__(self):
-        super().__init__()
-        self.default_seq_config_file = self._get_default_sequence_processing_config_file()
-        with open(self.default_seq_config_file, 'r') as file:
-            self.parameters = yaml.safe_load(file)
-
-        # Some postprocessing steps
-        self.parameters['tokenization']['shift']['constraints']['max'] = self.parameters['tokenization']['kmer']['default']-1
-        # Ha valaki update-li a k-mer paramter-t, akkor triggerelni kellene, hogy mi legyen. 
-
-        self.get_and_set_segmentation_parameters()
-        self.get_and_set_tokenization_parameters()
-        self.get_and_set_computational_parameters()
-
-    def _get_default_sequence_processing_config_file(self) -> str:
-        """
-        Retrieve the default sequence processing configuration file.
-
-        :return: Path to the configuration file.
-        :rtype: str
-        """
-        current_path = pathlib.Path(__file__).parent
-        prokbert_seq_config_file = join(current_path, 'configs', 'sequence_processing.yaml')
-        self.current_path = current_path
-
-        try:
-            # Attempt to read the environment variable
-            prokbert_seq_config_file = os.environ['SEQ_CONFIG_FILE']
-        except KeyError:
-            # Handle the case when the environment variable is not found
-            print("SEQ_CONFIG_FILE environment variable has not been set. Using default value: {0}".format(prokbert_seq_config_file))
-        return prokbert_seq_config_file
-
-    
-    def get_and_set_segmentation_parameters(self, parameters: dict = {}) -> dict:
-        """
-        Retrieve and validate the provided parameters for segmentation.
-
-        :param parameters: A dictionary of parameters to be validated.
-        :type parameters: dict
-        :return: A dictionary of validated segmentation parameters.
-        :rtype: dict
-        :raises ValueError: If an invalid segmentation parameter is provided.
-        """
-        segmentation_params = {k: self.get_parameter('segmentation', k) for k in self.parameters['segmentation']}
-
-        for param, param_value in parameters.items():
-            if param not in segmentation_params:
-                raise ValueError(f"The provided {param} is an INVALID segmentation parameter! The valid parameters are: {list(segmentation_params.keys())}")
-            self.validate('segmentation', param, param_value)
-            segmentation_params[param] = param_value
-        self.segmentation_params = segmentation_params
-
-
-        return segmentation_params
-
-
-    def get_and_set_tokenization_parameters(self, parameters: dict = {}) -> dict:
-        # Updating the other parameters if necesseary, i.e. if k-mer has-been changed, then the shift is updated and we run a parameter check at the end
-
-        tokenization_params = {k: self.get_parameter('tokenization', k) for k in self.parameters['tokenization']}
-        for param, param_value in parameters.items():
-            if param not in tokenization_params:
-                raise ValueError(f"The provided {param} is an INVALID tokenization parameter! The valid parameters are: {list(tokenization_params.keys())}")
-            self.validate('tokenization', param, param_value)
-            tokenization_params[param] = param_value
-
-        # Loading and check the vocab file. It is assumed that its ordered dictionary
-        vocabfile=tokenization_params['vocabfile']
-        act_kmer = tokenization_params['kmer']
-        if vocabfile=='auto':
-            print(self.current_path)
-            vocabfile_path = join(self.current_path, 'data/prokbert_vocabs/', f'prokbert-base-dna{act_kmer}', 'vocab.txt')
-            tokenization_params['vocabfile'] = vocabfile_path
-        else:
-            vocabfile_path = vocabfile
-        with open(vocabfile_path) as vocabfile_in:
-            vocabmap = {line.strip(): i for i, line in enumerate(vocabfile_in)}
-        tokenization_params['vocabmap'] = vocabmap
-
-        # Loading the vocab
-        self.tokenization_params = tokenization_params
-        return tokenization_params    
-
-    def get_and_set_computational_parameters(self, parameters: dict = {}) -> dict:
-        """ Reading and validating the computational paramters
-        """
-
-        computational_params = {k: self.get_parameter('computation', k) for k in self.parameters['computation']}
-        core_count = cpu_count()
-
-        if computational_params['cpu_cores_for_segmentation'] == -1:
-            computational_params['cpu_cores_for_segmentation'] = core_count
-
-        if computational_params['cpu_cores_for_tokenization'] == -1:
-            computational_params['cpu_cores_for_tokenization'] = core_count
-
-        
-
-        for param, param_value in parameters.items():
-            if param not in computational_params:
-                raise ValueError(f"The provided {param} is an INVALID computation parameter! The valid parameters are: {list(computational_params.keys())}")
-            self.validate('computation', param, param_value)
-            computational_params[param] = param_value
-
-        np_tokentype= SeqConfig.numpy_dtype_mapping[computational_params['numpy_token_integer_prec_byte']]
-        computational_params['np_tokentype'] = np_tokentype
-        self.computational_params = computational_params
-        return computational_params
-
-
-    def get_maximum_segment_length_from_token_count_from_params(self):
-        """Calculating the maximum length of the segment from the token count """
-        max_token_counts = self.tokenization_params['token_limit']
-        shift = self.tokenization_params['shift']
-        kmer = self.tokenization_params['kmer']
-        return self.get_maximum_segment_length_from_token_count(max_token_counts, shift, kmer)
-
-    def get_maximum_token_count_from_max_length_from_params(self):
-        """Calculating the maximum length of the segment from the token count """
-
-
-        max_segment_length = self.tokenization_params['max_segment_length']
-        shift = self.tokenization_params['shift']
-        kmer = self.tokenization_params['kmer']          
-        max_token_count = self.get_maximum_token_count_from_max_length(max_segment_length, shift, kmer)
-
-        return max_token_count
-
-    @staticmethod
-    def get_maximum_segment_length_from_token_count(max_token_counts, shift, kmer):
-        """Calcuates how long sequence can be covered
-        """
-
-        max_segment_length = (max_token_counts-3)*shift + kmer
-        return max_segment_length
-
-    @staticmethod
-    def get_maximum_token_count_from_max_length(max_segment_length, shift, kmer):
-        """Calcuates how long sequence can be covered
-        """
-        max_token_count = int(np.ceil((max_segment_length - kmer)/shift+3))
-        return max_token_count
-
-class ProkBERTConfig(BaseConfig):
-    """Class to manage and validate pretraining configurations."""
-
-    torch_dtype_mapping = {1: torch.uint8,
-                           2: torch.int16,
-                           8: torch.int64,
-                           4: torch.int32}
-
-    def __init__(self):
-        super().__init__()
-
-        self.default_pretrain_config_file = self._get_default_pretrain_config_file()
-        with open(self.default_pretrain_config_file, 'r') as file:
-            self.parameters = yaml.safe_load(file)
-            
-        # Load and validate each parameter set
-        self.data_collator_params = self.get_set_parameters('data_collator')
-        self.model_params = self.get_set_parameters('model')
-        self.dataset_params = self.get_set_parameters('dataset')
-        self.pretraining_params = self.get_set_parameters('pretraining')
-        # Getting the sequtils params as well
-
-        self.def_seq_config = SeqConfig()
-        self.segmentation_params = self.def_seq_config.get_and_set_segmentation_parameters(self.parameters['segmentation'])
-        self.tokenization_params = self.def_seq_config.get_and_set_tokenization_parameters(self.parameters['tokenization'])
-        self.computation_params = self.def_seq_config.get_and_set_computational_parameters(self.parameters['computation'])
-
-        self.default_torchtype = ProkBERTConfig.torch_dtype_mapping[self.computation_params['numpy_token_integer_prec_byte']]
-
-    def _get_default_pretrain_config_file(self) -> str:
-        """
-        Retrieve the default pretraining configuration file.
-
-        :return: Path to the configuration file.
-        :rtype: str
-        """
-        current_path = pathlib.Path(__file__).parent
-        pretrain_config_file = join(current_path, 'configs', 'pretraining.yaml')
-
-        try:
-            # Attempt to read the environment variable
-            pretrain_config_file = os.environ['PRETRAIN_CONFIG_FILE']
-        except KeyError:
-            # Handle the case when the environment variable is not found
-            print(f"PRETRAIN_CONFIG_FILE environment variable has not been set. Using default value: {pretrain_config_file}")
-        return pretrain_config_file
-    
-    def get_set_parameters(self, parameter_class: str, parameters: dict = {}) -> dict:
-        """
-        Retrieve and validate the provided parameters for a given parameter class.
-
-        :param parameter_class: The class/category of the parameter (e.g., 'data_collator').
-        :type parameter_class: str
-        :param parameters: A dictionary of parameters to be validated.
-        :type parameters: dict
-        :return: A dictionary of validated parameters.
-        :rtype: dict
-        :raises ValueError: If an invalid parameter is provided.
-        """
-        class_params = {k: self.get_parameter(parameter_class, k) for k in self.parameters[parameter_class]}
-
-        # First validatiading the class parameters as well
-        for param, param_value in class_params.items():
-
-            self.validate(parameter_class, param, param_value)
-
-
-        for param, param_value in parameters.items():
-            if param not in class_params:
-                raise ValueError(f"The provided {param} is an INVALID {parameter_class} parameter! The valid parameters are: {list(class_params.keys())}")
-            self.validate(parameter_class, param, param_value)
-            class_params[param] = param_value
-
-        return class_params
-    
-    def get_and_set_model_parameters(self, parameters: dict = {}) -> dict:
-        """ Setting the model parameters """
-
-        self.model_params = self.get_set_parameters('model', parameters)
-
-        return self.model_params
-
-    def get_and_set_dataset_parameters(self, parameters: dict = {}) -> dict:
-        """ Setting the dataset parameters """
-
-        self.dataset_params = self.get_set_parameters('dataset', parameters)
-
-        return self.dataset_params
-
-    def get_and_set_pretraining_parameters(self, parameters: dict = {}) -> dict:
-        """ Setting the model parameters """
-        self.pretraining_params = self.get_set_parameters('pretraining', parameters)
-
-        return self.pretraining_params       
-    
-    
-    def get_and_set_datacollator_parameters(self, parameters: dict = {}) -> dict:
-        """ Setting the model parameters """
-        self.data_collator_params = self.get_set_parameters('data_collator', parameters)
-        return self.data_collator_params
-    
-    def get_and_set_segmentation_parameters(self, parameters: dict = {}) -> dict:
-        self.segmentation_params = self.def_seq_config.get_and_set_segmentation_parameters(parameters)
-
-        return self.segmentation_params 
-    def get_and_set_tokenization_parameters(self, parameters: dict = {}) -> dict:
-        self.tokenization_params = self.def_seq_config.get_and_set_tokenization_parameters(parameters)
-        
-        return self.tokenization_params 
-    def get_and_set_computation_params(self, parameters: dict = {}) -> dict:
-        self.computation_params = self.def_seq_config.get_and_set_computational_parameters(parameters)
-        return self.computation_params    

data/prokbert_vocabs/prokbert-base-dna1/vocab.txt

@@ -1,9 +0,0 @@
-[PAD]
-[UNK]
-[CLS]
-[SEP]
-[MASK]
-A
-C
-G
-T

data/prokbert_vocabs/prokbert-base-dna2/vocab.txt

@@ -1,21 +0,0 @@
-[PAD]
-[UNK]
-[CLS]
-[SEP]
-[MASK]
-AA
-AC
-AG
-AT
-CA
-CC
-CG
-CT
-GA
-GC
-GG
-GT
-TA
-TC
-TG
-TT

data/prokbert_vocabs/prokbert-base-dna3/vocab.txt

@@ -1,69 +0,0 @@
-[PAD]
-[UNK]
-[CLS]
-[SEP]
-[MASK]
-AAA
-AAC
-AAG
-AAT
-ACA
-ACC
-ACG
-ACT
-AGA
-AGC
-AGG
-AGT
-ATA
-ATC
-ATG
-ATT
-CAA
-CAC
-CAG
-CAT
-CCA
-CCC
-CCG
-CCT
-CGA
-CGC
-CGG
-CGT
-CTA
-CTC
-CTG
-CTT
-GAA
-GAC
-GAG
-GAT
-GCA
-GCC
-GCG
-GCT
-GGA
-GGC
-GGG
-GGT
-GTA
-GTC
-GTG
-GTT
-TAA
-TAC
-TAG
-TAT
-TCA
-TCC
-TCG
-TCT
-TGA
-TGC
-TGG
-TGT
-TTA
-TTC
-TTG
-TTT

data/prokbert_vocabs/prokbert-base-dna4/vocab.txt

@@ -1,261 +0,0 @@
-[PAD]
-[UNK]
-[CLS]
-[SEP]
-[MASK]
-AAAA
-AAAC
-AAAG
-AAAT
-AACA
-AACC
-AACG
-AACT
-AAGA
-AAGC
-AAGG
-AAGT
-AATA
-AATC
-AATG
-AATT
-ACAA
-ACAC
-ACAG
-ACAT
-ACCA
-ACCC
-ACCG
-ACCT
-ACGA
-ACGC
-ACGG
-ACGT
-ACTA
-ACTC
-ACTG
-ACTT
-AGAA
-AGAC
-AGAG
-AGAT
-AGCA
-AGCC
-AGCG
-AGCT
-AGGA
-AGGC
-AGGG
-AGGT
-AGTA
-AGTC
-AGTG
-AGTT
-ATAA
-ATAC
-ATAG
-ATAT
-ATCA
-ATCC
-ATCG
-ATCT
-ATGA
-ATGC
-ATGG
-ATGT
-ATTA
-ATTC
-ATTG
-ATTT
-CAAA
-CAAC
-CAAG
-CAAT
-CACA
-CACC
-CACG
-CACT
-CAGA
-CAGC
-CAGG
-CAGT
-CATA
-CATC
-CATG
-CATT
-CCAA
-CCAC
-CCAG
-CCAT
-CCCA
-CCCC
-CCCG
-CCCT
-CCGA
-CCGC
-CCGG
-CCGT
-CCTA
-CCTC
-CCTG
-CCTT
-CGAA
-CGAC
-CGAG
-CGAT
-CGCA
-CGCC
-CGCG
-CGCT
-CGGA
-CGGC
-CGGG
-CGGT
-CGTA
-CGTC
-CGTG
-CGTT
-CTAA
-CTAC
-CTAG
-CTAT
-CTCA
-CTCC
-CTCG
-CTCT
-CTGA
-CTGC
-CTGG
-CTGT
-CTTA
-CTTC
-CTTG
-CTTT
-GAAA
-GAAC
-GAAG
-GAAT
-GACA
-GACC
-GACG
-GACT
-GAGA
-GAGC
-GAGG
-GAGT
-GATA
-GATC
-GATG
-GATT
-GCAA
-GCAC
-GCAG
-GCAT
-GCCA
-GCCC
-GCCG
-GCCT
-GCGA
-GCGC
-GCGG
-GCGT
-GCTA
-GCTC
-GCTG
-GCTT
-GGAA
-GGAC
-GGAG
-GGAT
-GGCA
-GGCC
-GGCG
-GGCT
-GGGA
-GGGC
-GGGG
-GGGT
-GGTA
-GGTC
-GGTG
-GGTT
-GTAA
-GTAC
-GTAG
-GTAT
-GTCA
-GTCC
-GTCG
-GTCT
-GTGA
-GTGC
-GTGG
-GTGT
-GTTA
-GTTC
-GTTG
-GTTT
-TAAA
-TAAC
-TAAG
-TAAT
-TACA
-TACC
-TACG
-TACT
-TAGA
-TAGC
-TAGG
-TAGT
-TATA
-TATC
-TATG
-TATT
-TCAA
-TCAC
-TCAG
-TCAT
-TCCA
-TCCC
-TCCG
-TCCT
-TCGA
-TCGC
-TCGG
-TCGT
-TCTA
-TCTC
-TCTG
-TCTT
-TGAA
-TGAC
-TGAG
-TGAT
-TGCA
-TGCC
-TGCG
-TGCT
-TGGA
-TGGC
-TGGG
-TGGT
-TGTA
-TGTC
-TGTG
-TGTT
-TTAA
-TTAC
-TTAG
-TTAT
-TTCA
-TTCC
-TTCG
-TTCT
-TTGA
-TTGC
-TTGG
-TTGT
-TTTA
-TTTC
-TTTG
-TTTT

data/prokbert_vocabs/prokbert-base-dna5/vocab.txt

@@ -1,1029 +0,0 @@
-[PAD]
-[UNK]
-[CLS]
-[SEP]
-[MASK]
-AAAAA
-AAAAC
-AAAAG
-AAAAT
-AAACA
-AAACC
-AAACG
-AAACT
-AAAGA
-AAAGC
-AAAGG
-AAAGT
-AAATA
-AAATC
-AAATG
-AAATT
-AACAA
-AACAC
-AACAG
-AACAT
-AACCA
-AACCC
-AACCG
-AACCT
-AACGA
-AACGC
-AACGG
-AACGT
-AACTA
-AACTC
-AACTG
-AACTT
-AAGAA
-AAGAC
-AAGAG
-AAGAT
-AAGCA
-AAGCC
-AAGCG
-AAGCT
-AAGGA
-AAGGC
-AAGGG
-AAGGT
-AAGTA
-AAGTC
-AAGTG
-AAGTT
-AATAA
-AATAC
-AATAG
-AATAT
-AATCA
-AATCC
-AATCG
-AATCT
-AATGA
-AATGC
-AATGG
-AATGT
-AATTA
-AATTC
-AATTG
-AATTT
-ACAAA
-ACAAC
-ACAAG
-ACAAT
-ACACA
-ACACC
-ACACG
-ACACT
-ACAGA
-ACAGC
-ACAGG
-ACAGT
-ACATA
-ACATC
-ACATG
-ACATT
-ACCAA
-ACCAC
-ACCAG
-ACCAT
-ACCCA
-ACCCC
-ACCCG
-ACCCT
-ACCGA
-ACCGC
-ACCGG
-ACCGT
-ACCTA
-ACCTC
-ACCTG
-ACCTT
-ACGAA
-ACGAC
-ACGAG
-ACGAT
-ACGCA
-ACGCC
-ACGCG
-ACGCT
-ACGGA
-ACGGC
-ACGGG
-ACGGT
-ACGTA
-ACGTC
-ACGTG
-ACGTT
-ACTAA
-ACTAC
-ACTAG
-ACTAT
-ACTCA
-ACTCC
-ACTCG
-ACTCT
-ACTGA
-ACTGC
-ACTGG
-ACTGT
-ACTTA
-ACTTC
-ACTTG
-ACTTT
-AGAAA
-AGAAC
-AGAAG
-AGAAT
-AGACA
-AGACC
-AGACG
-AGACT
-AGAGA
-AGAGC
-AGAGG
-AGAGT
-AGATA
-AGATC
-AGATG
-AGATT
-AGCAA
-AGCAC
-AGCAG
-AGCAT
-AGCCA
-AGCCC
-AGCCG
-AGCCT
-AGCGA
-AGCGC
-AGCGG
-AGCGT
-AGCTA
-AGCTC
-AGCTG
-AGCTT
-AGGAA
-AGGAC
-AGGAG
-AGGAT
-AGGCA
-AGGCC
-AGGCG
-AGGCT
-AGGGA
-AGGGC
-AGGGG
-AGGGT
-AGGTA
-AGGTC
-AGGTG
-AGGTT
-AGTAA
-AGTAC
-AGTAG
-AGTAT
-AGTCA
-AGTCC
-AGTCG
-AGTCT
-AGTGA
-AGTGC
-AGTGG
-AGTGT
-AGTTA
-AGTTC
-AGTTG
-AGTTT
-ATAAA
-ATAAC
-ATAAG
-ATAAT
-ATACA
-ATACC
-ATACG
-ATACT
-ATAGA
-ATAGC
-ATAGG
-ATAGT
-ATATA
-ATATC
-ATATG
-ATATT
-ATCAA
-ATCAC
-ATCAG
-ATCAT
-ATCCA
-ATCCC
-ATCCG
-ATCCT
-ATCGA
-ATCGC
-ATCGG
-ATCGT
-ATCTA
-ATCTC
-ATCTG
-ATCTT
-ATGAA
-ATGAC
-ATGAG
-ATGAT
-ATGCA
-ATGCC
-ATGCG
-ATGCT
-ATGGA
-ATGGC
-ATGGG
-ATGGT
-ATGTA
-ATGTC
-ATGTG
-ATGTT
-ATTAA
-ATTAC
-ATTAG
-ATTAT
-ATTCA
-ATTCC
-ATTCG
-ATTCT
-ATTGA
-ATTGC
-ATTGG
-ATTGT
-ATTTA
-ATTTC
-ATTTG
-ATTTT
-CAAAA
-CAAAC
-CAAAG
-CAAAT
-CAACA
-CAACC
-CAACG
-CAACT
-CAAGA
-CAAGC
-CAAGG
-CAAGT
-CAATA
-CAATC
-CAATG
-CAATT
-CACAA
-CACAC
-CACAG
-CACAT
-CACCA
-CACCC
-CACCG
-CACCT
-CACGA
-CACGC
-CACGG
-CACGT
-CACTA
-CACTC
-CACTG
-CACTT
-CAGAA
-CAGAC
-CAGAG
-CAGAT
-CAGCA
-CAGCC
-CAGCG
-CAGCT
-CAGGA
-CAGGC
-CAGGG
-CAGGT
-CAGTA
-CAGTC
-CAGTG
-CAGTT
-CATAA
-CATAC
-CATAG
-CATAT
-CATCA
-CATCC
-CATCG
-CATCT
-CATGA
-CATGC
-CATGG
-CATGT
-CATTA
-CATTC
-CATTG
-CATTT
-CCAAA
-CCAAC
-CCAAG
-CCAAT
-CCACA
-CCACC
-CCACG
-CCACT
-CCAGA
-CCAGC
-CCAGG
-CCAGT
-CCATA
-CCATC
-CCATG
-CCATT
-CCCAA
-CCCAC
-CCCAG
-CCCAT
-CCCCA
-CCCCC
-CCCCG
-CCCCT
-CCCGA
-CCCGC
-CCCGG
-CCCGT
-CCCTA
-CCCTC
-CCCTG
-CCCTT
-CCGAA
-CCGAC
-CCGAG
-CCGAT
-CCGCA
-CCGCC
-CCGCG
-CCGCT
-CCGGA
-CCGGC
-CCGGG
-CCGGT
-CCGTA
-CCGTC
-CCGTG
-CCGTT
-CCTAA
-CCTAC
-CCTAG
-CCTAT
-CCTCA
-CCTCC
-CCTCG
-CCTCT
-CCTGA
-CCTGC
-CCTGG
-CCTGT
-CCTTA
-CCTTC
-CCTTG
-CCTTT
-CGAAA
-CGAAC
-CGAAG
-CGAAT
-CGACA
-CGACC
-CGACG
-CGACT
-CGAGA
-CGAGC
-CGAGG
-CGAGT
-CGATA
-CGATC
-CGATG
-CGATT
-CGCAA
-CGCAC
-CGCAG
-CGCAT
-CGCCA
-CGCCC
-CGCCG
-CGCCT
-CGCGA
-CGCGC
-CGCGG
-CGCGT
-CGCTA
-CGCTC
-CGCTG
-CGCTT
-CGGAA
-CGGAC
-CGGAG
-CGGAT
-CGGCA
-CGGCC
-CGGCG
-CGGCT
-CGGGA
-CGGGC
-CGGGG
-CGGGT
-CGGTA
-CGGTC
-CGGTG
-CGGTT
-CGTAA
-CGTAC
-CGTAG
-CGTAT
-CGTCA
-CGTCC
-CGTCG
-CGTCT
-CGTGA
-CGTGC
-CGTGG
-CGTGT
-CGTTA
-CGTTC
-CGTTG
-CGTTT
-CTAAA
-CTAAC
-CTAAG
-CTAAT
-CTACA
-CTACC
-CTACG
-CTACT
-CTAGA
-CTAGC
-CTAGG
-CTAGT
-CTATA
-CTATC
-CTATG
-CTATT
-CTCAA
-CTCAC
-CTCAG
-CTCAT
-CTCCA
-CTCCC
-CTCCG
-CTCCT
-CTCGA
-CTCGC
-CTCGG
-CTCGT
-CTCTA
-CTCTC
-CTCTG
-CTCTT
-CTGAA
-CTGAC
-CTGAG
-CTGAT
-CTGCA
-CTGCC
-CTGCG
-CTGCT
-CTGGA
-CTGGC
-CTGGG
-CTGGT
-CTGTA
-CTGTC
-CTGTG
-CTGTT
-CTTAA
-CTTAC
-CTTAG
-CTTAT
-CTTCA
-CTTCC
-CTTCG
-CTTCT
-CTTGA
-CTTGC
-CTTGG
-CTTGT
-CTTTA
-CTTTC
-CTTTG
-CTTTT
-GAAAA
-GAAAC
-GAAAG
-GAAAT
-GAACA
-GAACC
-GAACG
-GAACT
-GAAGA
-GAAGC
-GAAGG
-GAAGT
-GAATA
-GAATC
-GAATG
-GAATT
-GACAA
-GACAC
-GACAG
-GACAT
-GACCA
-GACCC
-GACCG
-GACCT
-GACGA
-GACGC
-GACGG
-GACGT
-GACTA
-GACTC
-GACTG
-GACTT
-GAGAA
-GAGAC
-GAGAG
-GAGAT
-GAGCA
-GAGCC
-GAGCG
-GAGCT
-GAGGA
-GAGGC
-GAGGG
-GAGGT
-GAGTA
-GAGTC
-GAGTG
-GAGTT
-GATAA
-GATAC
-GATAG
-GATAT
-GATCA
-GATCC
-GATCG
-GATCT
-GATGA
-GATGC
-GATGG
-GATGT
-GATTA
-GATTC
-GATTG
-GATTT
-GCAAA
-GCAAC
-GCAAG
-GCAAT
-GCACA
-GCACC
-GCACG
-GCACT
-GCAGA
-GCAGC
-GCAGG
-GCAGT
-GCATA
-GCATC
-GCATG
-GCATT
-GCCAA
-GCCAC
-GCCAG
-GCCAT
-GCCCA
-GCCCC
-GCCCG
-GCCCT
-GCCGA
-GCCGC
-GCCGG
-GCCGT
-GCCTA
-GCCTC
-GCCTG
-GCCTT
-GCGAA
-GCGAC
-GCGAG
-GCGAT
-GCGCA
-GCGCC
-GCGCG
-GCGCT
-GCGGA
-GCGGC
-GCGGG
-GCGGT
-GCGTA
-GCGTC
-GCGTG
-GCGTT
-GCTAA
-GCTAC
-GCTAG
-GCTAT
-GCTCA
-GCTCC
-GCTCG
-GCTCT
-GCTGA
-GCTGC
-GCTGG
-GCTGT
-GCTTA
-GCTTC
-GCTTG
-GCTTT
-GGAAA
-GGAAC
-GGAAG
-GGAAT
-GGACA
-GGACC
-GGACG
-GGACT
-GGAGA
-GGAGC
-GGAGG
-GGAGT
-GGATA
-GGATC
-GGATG
-GGATT
-GGCAA
-GGCAC
-GGCAG
-GGCAT
-GGCCA
-GGCCC
-GGCCG
-GGCCT
-GGCGA
-GGCGC
-GGCGG
-GGCGT
-GGCTA
-GGCTC
-GGCTG
-GGCTT
-GGGAA
-GGGAC
-GGGAG
-GGGAT
-GGGCA
-GGGCC
-GGGCG
-GGGCT
-GGGGA
-GGGGC
-GGGGG
-GGGGT
-GGGTA
-GGGTC
-GGGTG
-GGGTT
-GGTAA
-GGTAC
-GGTAG
-GGTAT
-GGTCA
-GGTCC
-GGTCG
-GGTCT
-GGTGA
-GGTGC
-GGTGG
-GGTGT
-GGTTA
-GGTTC
-GGTTG
-GGTTT
-GTAAA
-GTAAC
-GTAAG
-GTAAT
-GTACA
-GTACC
-GTACG
-GTACT
-GTAGA
-GTAGC
-GTAGG
-GTAGT
-GTATA
-GTATC
-GTATG
-GTATT
-GTCAA
-GTCAC
-GTCAG
-GTCAT
-GTCCA
-GTCCC
-GTCCG
-GTCCT
-GTCGA
-GTCGC
-GTCGG
-GTCGT
-GTCTA
-GTCTC
-GTCTG
-GTCTT
-GTGAA
-GTGAC
-GTGAG
-GTGAT
-GTGCA
-GTGCC
-GTGCG
-GTGCT
-GTGGA
-GTGGC
-GTGGG
-GTGGT
-GTGTA
-GTGTC
-GTGTG
-GTGTT
-GTTAA
-GTTAC
-GTTAG
-GTTAT
-GTTCA
-GTTCC
-GTTCG
-GTTCT
-GTTGA
-GTTGC
-GTTGG
-GTTGT
-GTTTA
-GTTTC
-GTTTG
-GTTTT
-TAAAA
-TAAAC
-TAAAG
-TAAAT
-TAACA
-TAACC
-TAACG
-TAACT
-TAAGA
-TAAGC
-TAAGG
-TAAGT
-TAATA
-TAATC
-TAATG
-TAATT
-TACAA
-TACAC
-TACAG
-TACAT
-TACCA
-TACCC
-TACCG
-TACCT
-TACGA
-TACGC
-TACGG
-TACGT
-TACTA
-TACTC
-TACTG
-TACTT
-TAGAA
-TAGAC
-TAGAG
-TAGAT
-TAGCA
-TAGCC
-TAGCG
-TAGCT
-TAGGA
-TAGGC
-TAGGG
-TAGGT
-TAGTA
-TAGTC
-TAGTG
-TAGTT
-TATAA
-TATAC
-TATAG
-TATAT
-TATCA
-TATCC
-TATCG
-TATCT
-TATGA
-TATGC
-TATGG
-TATGT
-TATTA
-TATTC
-TATTG
-TATTT
-TCAAA
-TCAAC
-TCAAG
-TCAAT
-TCACA
-TCACC
-TCACG
-TCACT
-TCAGA
-TCAGC
-TCAGG
-TCAGT
-TCATA
-TCATC
-TCATG
-TCATT
-TCCAA
-TCCAC
-TCCAG
-TCCAT
-TCCCA
-TCCCC
-TCCCG
-TCCCT
-TCCGA
-TCCGC
-TCCGG
-TCCGT
-TCCTA
-TCCTC
-TCCTG
-TCCTT
-TCGAA
-TCGAC
-TCGAG
-TCGAT
-TCGCA
-TCGCC
-TCGCG
-TCGCT
-TCGGA
-TCGGC
-TCGGG
-TCGGT
-TCGTA
-TCGTC
-TCGTG
-TCGTT
-TCTAA
-TCTAC
-TCTAG
-TCTAT
-TCTCA
-TCTCC
-TCTCG
-TCTCT
-TCTGA
-TCTGC
-TCTGG
-TCTGT
-TCTTA
-TCTTC
-TCTTG
-TCTTT
-TGAAA
-TGAAC
-TGAAG
-TGAAT
-TGACA
-TGACC
-TGACG
-TGACT
-TGAGA
-TGAGC
-TGAGG
-TGAGT
-TGATA
-TGATC
-TGATG
-TGATT
-TGCAA
-TGCAC
-TGCAG
-TGCAT
-TGCCA
-TGCCC
-TGCCG
-TGCCT
-TGCGA
-TGCGC
-TGCGG
-TGCGT
-TGCTA
-TGCTC
-TGCTG
-TGCTT
-TGGAA
-TGGAC
-TGGAG
-TGGAT
-TGGCA
-TGGCC
-TGGCG
-TGGCT
-TGGGA
-TGGGC
-TGGGG
-TGGGT
-TGGTA
-TGGTC
-TGGTG
-TGGTT
-TGTAA
-TGTAC
-TGTAG
-TGTAT
-TGTCA
-TGTCC
-TGTCG
-TGTCT
-TGTGA
-TGTGC
-TGTGG
-TGTGT
-TGTTA
-TGTTC
-TGTTG
-TGTTT
-TTAAA
-TTAAC
-TTAAG
-TTAAT
-TTACA
-TTACC
-TTACG
-TTACT
-TTAGA
-TTAGC
-TTAGG
-TTAGT
-TTATA
-TTATC
-TTATG
-TTATT
-TTCAA
-TTCAC
-TTCAG
-TTCAT
-TTCCA
-TTCCC
-TTCCG
-TTCCT
-TTCGA
-TTCGC
-TTCGG
-TTCGT
-TTCTA
-TTCTC
-TTCTG
-TTCTT
-TTGAA
-TTGAC
-TTGAG
-TTGAT
-TTGCA
-TTGCC
-TTGCG
-TTGCT
-TTGGA
-TTGGC
-TTGGG
-TTGGT
-TTGTA
-TTGTC
-TTGTG
-TTGTT
-TTTAA
-TTTAC
-TTTAG
-TTTAT
-TTTCA
-TTTCC
-TTTCG
-TTTCT
-TTTGA
-TTTGC
-TTTGG
-TTTGT
-TTTTA
-TTTTC
-TTTTG
-TTTTT

data/prokbert_vocabs/prokbert-base-dna6/vocab.txt

@@ -1,4101 +0,0 @@
-[PAD]
-[UNK]
-[CLS]
-[SEP]
-[MASK]
-AAAAAA
-AAAAAC
-AAAAAG
-AAAAAT
-AAAACA
-AAAACC
-AAAACG
-AAAACT
-AAAAGA
-AAAAGC
-AAAAGG
-AAAAGT
-AAAATA
-AAAATC
-AAAATG
-AAAATT
-AAACAA
-AAACAC
-AAACAG
-AAACAT
-AAACCA
-AAACCC
-AAACCG
-AAACCT
-AAACGA
-AAACGC
-AAACGG
-AAACGT
-AAACTA
-AAACTC
-AAACTG
-AAACTT
-AAAGAA
-AAAGAC
-AAAGAG
-AAAGAT
-AAAGCA
-AAAGCC
-AAAGCG
-AAAGCT
-AAAGGA
-AAAGGC
-AAAGGG
-AAAGGT
-AAAGTA
-AAAGTC
-AAAGTG
-AAAGTT
-AAATAA
-AAATAC
-AAATAG
-AAATAT
-AAATCA
-AAATCC
-AAATCG
-AAATCT
-AAATGA
-AAATGC
-AAATGG
-AAATGT
-AAATTA
-AAATTC
-AAATTG
-AAATTT
-AACAAA
-AACAAC
-AACAAG
-AACAAT
-AACACA
-AACACC
-AACACG
-AACACT
-AACAGA
-AACAGC
-AACAGG
-AACAGT
-AACATA
-AACATC
-AACATG
-AACATT
-AACCAA
-AACCAC
-AACCAG
-AACCAT
-AACCCA
-AACCCC
-AACCCG
-AACCCT
-AACCGA
-AACCGC
-AACCGG
-AACCGT
-AACCTA
-AACCTC
-AACCTG
-AACCTT
-AACGAA
-AACGAC
-AACGAG
-AACGAT
-AACGCA
-AACGCC
-AACGCG
-AACGCT
-AACGGA
-AACGGC
-AACGGG
-AACGGT
-AACGTA
-AACGTC
-AACGTG
-AACGTT
-AACTAA
-AACTAC
-AACTAG
-AACTAT
-AACTCA
-AACTCC
-AACTCG
-AACTCT
-AACTGA
-AACTGC
-AACTGG
-AACTGT
-AACTTA
-AACTTC
-AACTTG
-AACTTT
-AAGAAA
-AAGAAC
-AAGAAG
-AAGAAT
-AAGACA
-AAGACC
-AAGACG
-AAGACT
-AAGAGA
-AAGAGC
-AAGAGG
-AAGAGT
-AAGATA
-AAGATC
-AAGATG
-AAGATT
-AAGCAA
-AAGCAC
-AAGCAG
-AAGCAT
-AAGCCA
-AAGCCC
-AAGCCG
-AAGCCT
-AAGCGA
-AAGCGC
-AAGCGG
-AAGCGT
-AAGCTA
-AAGCTC
-AAGCTG
-AAGCTT
-AAGGAA
-AAGGAC
-AAGGAG
-AAGGAT
-AAGGCA
-AAGGCC
-AAGGCG
-AAGGCT
-AAGGGA
-AAGGGC
-AAGGGG
-AAGGGT
-AAGGTA
-AAGGTC
-AAGGTG
-AAGGTT
-AAGTAA
-AAGTAC
-AAGTAG
-AAGTAT
-AAGTCA
-AAGTCC
-AAGTCG
-AAGTCT
-AAGTGA
-AAGTGC
-AAGTGG
-AAGTGT
-AAGTTA
-AAGTTC
-AAGTTG
-AAGTTT
-AATAAA
-AATAAC
-AATAAG
-AATAAT
-AATACA
-AATACC
-AATACG
-AATACT
-AATAGA
-AATAGC
-AATAGG
-AATAGT
-AATATA
-AATATC
-AATATG
-AATATT
-AATCAA
-AATCAC
-AATCAG
-AATCAT
-AATCCA
-AATCCC
-AATCCG
-AATCCT
-AATCGA
-AATCGC
-AATCGG
-AATCGT
-AATCTA
-AATCTC
-AATCTG
-AATCTT
-AATGAA
-AATGAC
-AATGAG
-AATGAT
-AATGCA
-AATGCC
-AATGCG
-AATGCT
-AATGGA
-AATGGC
-AATGGG
-AATGGT
-AATGTA
-AATGTC
-AATGTG
-AATGTT
-AATTAA
-AATTAC
-AATTAG
-AATTAT
-AATTCA
-AATTCC
-AATTCG
-AATTCT
-AATTGA
-AATTGC
-AATTGG
-AATTGT
-AATTTA
-AATTTC
-AATTTG
-AATTTT
-ACAAAA
-ACAAAC
-ACAAAG
-ACAAAT
-ACAACA
-ACAACC
-ACAACG
-ACAACT
-ACAAGA
-ACAAGC
-ACAAGG
-ACAAGT
-ACAATA
-ACAATC
-ACAATG
-ACAATT
-ACACAA
-ACACAC
-ACACAG
-ACACAT
-ACACCA
-ACACCC
-ACACCG
-ACACCT
-ACACGA
-ACACGC
-ACACGG
-ACACGT
-ACACTA
-ACACTC
-ACACTG
-ACACTT
-ACAGAA
-ACAGAC
-ACAGAG
-ACAGAT
-ACAGCA
-ACAGCC
-ACAGCG
-ACAGCT
-ACAGGA
-ACAGGC
-ACAGGG
-ACAGGT
-ACAGTA
-ACAGTC
-ACAGTG
-ACAGTT
-ACATAA
-ACATAC
-ACATAG
-ACATAT
-ACATCA
-ACATCC
-ACATCG
-ACATCT
-ACATGA
-ACATGC
-ACATGG
-ACATGT
-ACATTA
-ACATTC
-ACATTG
-ACATTT
-ACCAAA
-ACCAAC
-ACCAAG
-ACCAAT
-ACCACA
-ACCACC
-ACCACG
-ACCACT
-ACCAGA
-ACCAGC
-ACCAGG
-ACCAGT
-ACCATA
-ACCATC
-ACCATG
-ACCATT
-ACCCAA
-ACCCAC
-ACCCAG
-ACCCAT
-ACCCCA
-ACCCCC
-ACCCCG
-ACCCCT
-ACCCGA
-ACCCGC
-ACCCGG
-ACCCGT
-ACCCTA
-ACCCTC
-ACCCTG
-ACCCTT
-ACCGAA
-ACCGAC
-ACCGAG
-ACCGAT
-ACCGCA
-ACCGCC
-ACCGCG
-ACCGCT
-ACCGGA
-ACCGGC
-ACCGGG
-ACCGGT
-ACCGTA
-ACCGTC
-ACCGTG
-ACCGTT
-ACCTAA
-ACCTAC
-ACCTAG
-ACCTAT
-ACCTCA
-ACCTCC
-ACCTCG
-ACCTCT
-ACCTGA
-ACCTGC
-ACCTGG
-ACCTGT
-ACCTTA
-ACCTTC
-ACCTTG
-ACCTTT
-ACGAAA
-ACGAAC
-ACGAAG
-ACGAAT
-ACGACA
-ACGACC
-ACGACG
-ACGACT
-ACGAGA
-ACGAGC
-ACGAGG
-ACGAGT
-ACGATA
-ACGATC
-ACGATG
-ACGATT
-ACGCAA
-ACGCAC
-ACGCAG
-ACGCAT
-ACGCCA
-ACGCCC
-ACGCCG
-ACGCCT
-ACGCGA
-ACGCGC
-ACGCGG
-ACGCGT
-ACGCTA
-ACGCTC
-ACGCTG
-ACGCTT
-ACGGAA
-ACGGAC
-ACGGAG
-ACGGAT
-ACGGCA
-ACGGCC
-ACGGCG
-ACGGCT
-ACGGGA
-ACGGGC
-ACGGGG
-ACGGGT
-ACGGTA
-ACGGTC
-ACGGTG
-ACGGTT
-ACGTAA
-ACGTAC
-ACGTAG
-ACGTAT
-ACGTCA
-ACGTCC
-ACGTCG
-ACGTCT
-ACGTGA
-ACGTGC
-ACGTGG
-ACGTGT
-ACGTTA
-ACGTTC
-ACGTTG
-ACGTTT
-ACTAAA
-ACTAAC
-ACTAAG
-ACTAAT
-ACTACA
-ACTACC
-ACTACG
-ACTACT
-ACTAGA
-ACTAGC
-ACTAGG
-ACTAGT
-ACTATA
-ACTATC
-ACTATG
-ACTATT
-ACTCAA
-ACTCAC
-ACTCAG
-ACTCAT
-ACTCCA
-ACTCCC
-ACTCCG
-ACTCCT
-ACTCGA
-ACTCGC
-ACTCGG
-ACTCGT
-ACTCTA
-ACTCTC
-ACTCTG
-ACTCTT
-ACTGAA
-ACTGAC
-ACTGAG
-ACTGAT
-ACTGCA
-ACTGCC
-ACTGCG
-ACTGCT
-ACTGGA
-ACTGGC
-ACTGGG
-ACTGGT
-ACTGTA
-ACTGTC
-ACTGTG
-ACTGTT
-ACTTAA
-ACTTAC
-ACTTAG
-ACTTAT
-ACTTCA
-ACTTCC
-ACTTCG
-ACTTCT
-ACTTGA
-ACTTGC
-ACTTGG
-ACTTGT
-ACTTTA
-ACTTTC
-ACTTTG
-ACTTTT
-AGAAAA
-AGAAAC
-AGAAAG
-AGAAAT
-AGAACA
-AGAACC
-AGAACG
-AGAACT
-AGAAGA
-AGAAGC
-AGAAGG
-AGAAGT
-AGAATA
-AGAATC
-AGAATG
-AGAATT
-AGACAA
-AGACAC
-AGACAG
-AGACAT
-AGACCA
-AGACCC
-AGACCG
-AGACCT
-AGACGA
-AGACGC
-AGACGG
-AGACGT
-AGACTA
-AGACTC
-AGACTG
-AGACTT
-AGAGAA
-AGAGAC
-AGAGAG
-AGAGAT
-AGAGCA
-AGAGCC
-AGAGCG
-AGAGCT
-AGAGGA
-AGAGGC
-AGAGGG
-AGAGGT
-AGAGTA
-AGAGTC
-AGAGTG
-AGAGTT
-AGATAA
-AGATAC
-AGATAG
-AGATAT
-AGATCA
-AGATCC
-AGATCG
-AGATCT
-AGATGA
-AGATGC
-AGATGG
-AGATGT
-AGATTA
-AGATTC
-AGATTG
-AGATTT
-AGCAAA
-AGCAAC
-AGCAAG
-AGCAAT
-AGCACA
-AGCACC
-AGCACG
-AGCACT
-AGCAGA
-AGCAGC
-AGCAGG
-AGCAGT
-AGCATA
-AGCATC
-AGCATG
-AGCATT
-AGCCAA
-AGCCAC
-AGCCAG
-AGCCAT
-AGCCCA
-AGCCCC
-AGCCCG
-AGCCCT
-AGCCGA
-AGCCGC
-AGCCGG
-AGCCGT
-AGCCTA
-AGCCTC
-AGCCTG
-AGCCTT
-AGCGAA
-AGCGAC
-AGCGAG
-AGCGAT
-AGCGCA
-AGCGCC
-AGCGCG
-AGCGCT
-AGCGGA
-AGCGGC
-AGCGGG
-AGCGGT
-AGCGTA
-AGCGTC
-AGCGTG
-AGCGTT
-AGCTAA
-AGCTAC
-AGCTAG
-AGCTAT
-AGCTCA
-AGCTCC
-AGCTCG
-AGCTCT
-AGCTGA
-AGCTGC
-AGCTGG
-AGCTGT
-AGCTTA
-AGCTTC
-AGCTTG
-AGCTTT
-AGGAAA
-AGGAAC
-AGGAAG
-AGGAAT
-AGGACA
-AGGACC
-AGGACG
-AGGACT
-AGGAGA
-AGGAGC
-AGGAGG
-AGGAGT
-AGGATA
-AGGATC
-AGGATG
-AGGATT
-AGGCAA
-AGGCAC
-AGGCAG
-AGGCAT
-AGGCCA
-AGGCCC
-AGGCCG
-AGGCCT
-AGGCGA
-AGGCGC
-AGGCGG
-AGGCGT
-AGGCTA
-AGGCTC
-AGGCTG
-AGGCTT
-AGGGAA
-AGGGAC
-AGGGAG
-AGGGAT
-AGGGCA
-AGGGCC
-AGGGCG
-AGGGCT
-AGGGGA
-AGGGGC
-AGGGGG
-AGGGGT
-AGGGTA
-AGGGTC
-AGGGTG
-AGGGTT
-AGGTAA
-AGGTAC
-AGGTAG
-AGGTAT
-AGGTCA
-AGGTCC
-AGGTCG
-AGGTCT
-AGGTGA
-AGGTGC
-AGGTGG
-AGGTGT
-AGGTTA
-AGGTTC
-AGGTTG
-AGGTTT
-AGTAAA
-AGTAAC
-AGTAAG
-AGTAAT
-AGTACA
-AGTACC
-AGTACG
-AGTACT
-AGTAGA
-AGTAGC
-AGTAGG
-AGTAGT
-AGTATA
-AGTATC
-AGTATG
-AGTATT
-AGTCAA
-AGTCAC
-AGTCAG
-AGTCAT
-AGTCCA
-AGTCCC
-AGTCCG
-AGTCCT
-AGTCGA
-AGTCGC
-AGTCGG
-AGTCGT
-AGTCTA
-AGTCTC
-AGTCTG
-AGTCTT
-AGTGAA
-AGTGAC
-AGTGAG
-AGTGAT
-AGTGCA
-AGTGCC
-AGTGCG
-AGTGCT
-AGTGGA
-AGTGGC
-AGTGGG
-AGTGGT
-AGTGTA
-AGTGTC
-AGTGTG
-AGTGTT
-AGTTAA
-AGTTAC
-AGTTAG
-AGTTAT
-AGTTCA
-AGTTCC
-AGTTCG
-AGTTCT
-AGTTGA
-AGTTGC
-AGTTGG
-AGTTGT
-AGTTTA
-AGTTTC
-AGTTTG
-AGTTTT
-ATAAAA
-ATAAAC
-ATAAAG
-ATAAAT
-ATAACA
-ATAACC
-ATAACG
-ATAACT
-ATAAGA
-ATAAGC
-ATAAGG
-ATAAGT
-ATAATA
-ATAATC
-ATAATG
-ATAATT
-ATACAA
-ATACAC
-ATACAG
-ATACAT
-ATACCA
-ATACCC
-ATACCG
-ATACCT
-ATACGA
-ATACGC
-ATACGG
-ATACGT
-ATACTA
-ATACTC
-ATACTG
-ATACTT
-ATAGAA
-ATAGAC
-ATAGAG
-ATAGAT
-ATAGCA
-ATAGCC
-ATAGCG
-ATAGCT
-ATAGGA
-ATAGGC
-ATAGGG
-ATAGGT
-ATAGTA
-ATAGTC
-ATAGTG
-ATAGTT
-ATATAA
-ATATAC
-ATATAG
-ATATAT
-ATATCA
-ATATCC
-ATATCG
-ATATCT
-ATATGA
-ATATGC
-ATATGG
-ATATGT
-ATATTA
-ATATTC
-ATATTG
-ATATTT
-ATCAAA
-ATCAAC
-ATCAAG
-ATCAAT
-ATCACA
-ATCACC
-ATCACG
-ATCACT
-ATCAGA
-ATCAGC
-ATCAGG
-ATCAGT
-ATCATA
-ATCATC
-ATCATG
-ATCATT
-ATCCAA
-ATCCAC
-ATCCAG
-ATCCAT
-ATCCCA
-ATCCCC
-ATCCCG
-ATCCCT
-ATCCGA
-ATCCGC
-ATCCGG
-ATCCGT
-ATCCTA
-ATCCTC
-ATCCTG
-ATCCTT
-ATCGAA
-ATCGAC
-ATCGAG
-ATCGAT
-ATCGCA
-ATCGCC
-ATCGCG
-ATCGCT
-ATCGGA
-ATCGGC
-ATCGGG
-ATCGGT
-ATCGTA
-ATCGTC
-ATCGTG
-ATCGTT
-ATCTAA
-ATCTAC
-ATCTAG
-ATCTAT
-ATCTCA
-ATCTCC
-ATCTCG
-ATCTCT
-ATCTGA
-ATCTGC
-ATCTGG
-ATCTGT
-ATCTTA
-ATCTTC
-ATCTTG
-ATCTTT
-ATGAAA
-ATGAAC
-ATGAAG
-ATGAAT
-ATGACA
-ATGACC
-ATGACG
-ATGACT
-ATGAGA
-ATGAGC
-ATGAGG
-ATGAGT
-ATGATA
-ATGATC
-ATGATG
-ATGATT
-ATGCAA
-ATGCAC
-ATGCAG
-ATGCAT
-ATGCCA
-ATGCCC
-ATGCCG
-ATGCCT
-ATGCGA
-ATGCGC
-ATGCGG
-ATGCGT
-ATGCTA
-ATGCTC
-ATGCTG
-ATGCTT
-ATGGAA
-ATGGAC
-ATGGAG
-ATGGAT
-ATGGCA
-ATGGCC
-ATGGCG
-ATGGCT
-ATGGGA
-ATGGGC
-ATGGGG
-ATGGGT
-ATGGTA
-ATGGTC
-ATGGTG
-ATGGTT
-ATGTAA
-ATGTAC
-ATGTAG
-ATGTAT
-ATGTCA
-ATGTCC
-ATGTCG
-ATGTCT
-ATGTGA
-ATGTGC
-ATGTGG
-ATGTGT
-ATGTTA
-ATGTTC
-ATGTTG
-ATGTTT
-ATTAAA
-ATTAAC
-ATTAAG
-ATTAAT
-ATTACA
-ATTACC
-ATTACG
-ATTACT
-ATTAGA
-ATTAGC
-ATTAGG
-ATTAGT
-ATTATA
-ATTATC
-ATTATG
-ATTATT
-ATTCAA
-ATTCAC
-ATTCAG
-ATTCAT
-ATTCCA
-ATTCCC
-ATTCCG
-ATTCCT
-ATTCGA
-ATTCGC
-ATTCGG
-ATTCGT
-ATTCTA
-ATTCTC
-ATTCTG
-ATTCTT
-ATTGAA
-ATTGAC
-ATTGAG
-ATTGAT
-ATTGCA
-ATTGCC
-ATTGCG
-ATTGCT
-ATTGGA
-ATTGGC
-ATTGGG
-ATTGGT
-ATTGTA
-ATTGTC
-ATTGTG
-ATTGTT
-ATTTAA
-ATTTAC
-ATTTAG
-ATTTAT
-ATTTCA
-ATTTCC
-ATTTCG
-ATTTCT
-ATTTGA
-ATTTGC
-ATTTGG
-ATTTGT
-ATTTTA
-ATTTTC
-ATTTTG
-ATTTTT
-CAAAAA
-CAAAAC
-CAAAAG
-CAAAAT
-CAAACA
-CAAACC
-CAAACG
-CAAACT
-CAAAGA
-CAAAGC
-CAAAGG
-CAAAGT
-CAAATA
-CAAATC
-CAAATG
-CAAATT
-CAACAA
-CAACAC
-CAACAG
-CAACAT
-CAACCA
-CAACCC
-CAACCG
-CAACCT
-CAACGA
-CAACGC
-CAACGG
-CAACGT
-CAACTA
-CAACTC
-CAACTG
-CAACTT
-CAAGAA
-CAAGAC
-CAAGAG
-CAAGAT
-CAAGCA
-CAAGCC
-CAAGCG
-CAAGCT
-CAAGGA
-CAAGGC
-CAAGGG
-CAAGGT
-CAAGTA
-CAAGTC
-CAAGTG
-CAAGTT
-CAATAA
-CAATAC
-CAATAG
-CAATAT
-CAATCA
-CAATCC
-CAATCG
-CAATCT
-CAATGA
-CAATGC
-CAATGG
-CAATGT
-CAATTA
-CAATTC
-CAATTG
-CAATTT
-CACAAA
-CACAAC
-CACAAG
-CACAAT
-CACACA
-CACACC
-CACACG
-CACACT
-CACAGA
-CACAGC
-CACAGG
-CACAGT
-CACATA
-CACATC
-CACATG
-CACATT
-CACCAA
-CACCAC
-CACCAG
-CACCAT
-CACCCA
-CACCCC
-CACCCG
-CACCCT
-CACCGA
-CACCGC
-CACCGG
-CACCGT
-CACCTA
-CACCTC
-CACCTG
-CACCTT
-CACGAA
-CACGAC
-CACGAG
-CACGAT
-CACGCA
-CACGCC
-CACGCG
-CACGCT
-CACGGA
-CACGGC
-CACGGG
-CACGGT
-CACGTA
-CACGTC
-CACGTG
-CACGTT
-CACTAA
-CACTAC
-CACTAG
-CACTAT
-CACTCA
-CACTCC
-CACTCG
-CACTCT
-CACTGA
-CACTGC
-CACTGG
-CACTGT
-CACTTA
-CACTTC
-CACTTG
-CACTTT
-CAGAAA
-CAGAAC
-CAGAAG
-CAGAAT
-CAGACA
-CAGACC
-CAGACG
-CAGACT
-CAGAGA
-CAGAGC
-CAGAGG
-CAGAGT
-CAGATA
-CAGATC
-CAGATG
-CAGATT
-CAGCAA
-CAGCAC
-CAGCAG
-CAGCAT
-CAGCCA
-CAGCCC
-CAGCCG
-CAGCCT
-CAGCGA
-CAGCGC
-CAGCGG
-CAGCGT
-CAGCTA
-CAGCTC
-CAGCTG
-CAGCTT
-CAGGAA
-CAGGAC
-CAGGAG
-CAGGAT
-CAGGCA
-CAGGCC
-CAGGCG
-CAGGCT
-CAGGGA
-CAGGGC
-CAGGGG
-CAGGGT
-CAGGTA
-CAGGTC
-CAGGTG
-CAGGTT
-CAGTAA
-CAGTAC
-CAGTAG
-CAGTAT
-CAGTCA
-CAGTCC
-CAGTCG
-CAGTCT
-CAGTGA
-CAGTGC
-CAGTGG
-CAGTGT
-CAGTTA
-CAGTTC
-CAGTTG
-CAGTTT
-CATAAA
-CATAAC
-CATAAG
-CATAAT
-CATACA
-CATACC
-CATACG
-CATACT
-CATAGA
-CATAGC
-CATAGG
-CATAGT
-CATATA
-CATATC
-CATATG
-CATATT
-CATCAA
-CATCAC
-CATCAG
-CATCAT
-CATCCA
-CATCCC
-CATCCG
-CATCCT
-CATCGA
-CATCGC
-CATCGG
-CATCGT
-CATCTA
-CATCTC
-CATCTG
-CATCTT
-CATGAA
-CATGAC
-CATGAG
-CATGAT
-CATGCA
-CATGCC
-CATGCG
-CATGCT
-CATGGA
-CATGGC
-CATGGG
-CATGGT
-CATGTA
-CATGTC
-CATGTG
-CATGTT
-CATTAA
-CATTAC
-CATTAG
-CATTAT
-CATTCA
-CATTCC
-CATTCG
-CATTCT
-CATTGA
-CATTGC
-CATTGG
-CATTGT
-CATTTA
-CATTTC
-CATTTG
-CATTTT
-CCAAAA
-CCAAAC
-CCAAAG
-CCAAAT
-CCAACA
-CCAACC
-CCAACG
-CCAACT
-CCAAGA
-CCAAGC
-CCAAGG
-CCAAGT
-CCAATA
-CCAATC
-CCAATG
-CCAATT
-CCACAA
-CCACAC
-CCACAG
-CCACAT
-CCACCA
-CCACCC
-CCACCG
-CCACCT
-CCACGA
-CCACGC
-CCACGG
-CCACGT
-CCACTA
-CCACTC
-CCACTG
-CCACTT
-CCAGAA
-CCAGAC
-CCAGAG
-CCAGAT
-CCAGCA
-CCAGCC
-CCAGCG
-CCAGCT
-CCAGGA
-CCAGGC
-CCAGGG
-CCAGGT
-CCAGTA
-CCAGTC
-CCAGTG
-CCAGTT
-CCATAA
-CCATAC
-CCATAG
-CCATAT
-CCATCA
-CCATCC
-CCATCG
-CCATCT
-CCATGA
-CCATGC
-CCATGG
-CCATGT
-CCATTA
-CCATTC
-CCATTG
-CCATTT
-CCCAAA
-CCCAAC
-CCCAAG
-CCCAAT
-CCCACA
-CCCACC
-CCCACG
-CCCACT
-CCCAGA
-CCCAGC
-CCCAGG
-CCCAGT
-CCCATA
-CCCATC
-CCCATG
-CCCATT
-CCCCAA
-CCCCAC
-CCCCAG
-CCCCAT
-CCCCCA
-CCCCCC
-CCCCCG
-CCCCCT
-CCCCGA
-CCCCGC
-CCCCGG
-CCCCGT
-CCCCTA
-CCCCTC
-CCCCTG
-CCCCTT
-CCCGAA
-CCCGAC
-CCCGAG
-CCCGAT
-CCCGCA
-CCCGCC
-CCCGCG
-CCCGCT
-CCCGGA
-CCCGGC
-CCCGGG
-CCCGGT
-CCCGTA
-CCCGTC
-CCCGTG
-CCCGTT
-CCCTAA
-CCCTAC
-CCCTAG
-CCCTAT
-CCCTCA
-CCCTCC
-CCCTCG
-CCCTCT
-CCCTGA
-CCCTGC
-CCCTGG
-CCCTGT
-CCCTTA
-CCCTTC
-CCCTTG
-CCCTTT
-CCGAAA
-CCGAAC
-CCGAAG
-CCGAAT
-CCGACA
-CCGACC
-CCGACG
-CCGACT
-CCGAGA
-CCGAGC
-CCGAGG
-CCGAGT
-CCGATA
-CCGATC
-CCGATG
-CCGATT
-CCGCAA
-CCGCAC
-CCGCAG
-CCGCAT
-CCGCCA
-CCGCCC
-CCGCCG
-CCGCCT
-CCGCGA
-CCGCGC
-CCGCGG
-CCGCGT
-CCGCTA
-CCGCTC
-CCGCTG
-CCGCTT
-CCGGAA
-CCGGAC
-CCGGAG
-CCGGAT
-CCGGCA
-CCGGCC
-CCGGCG
-CCGGCT
-CCGGGA
-CCGGGC
-CCGGGG
-CCGGGT
-CCGGTA
-CCGGTC
-CCGGTG
-CCGGTT
-CCGTAA
-CCGTAC
-CCGTAG
-CCGTAT
-CCGTCA
-CCGTCC
-CCGTCG
-CCGTCT
-CCGTGA
-CCGTGC
-CCGTGG
-CCGTGT
-CCGTTA
-CCGTTC
-CCGTTG
-CCGTTT
-CCTAAA
-CCTAAC
-CCTAAG
-CCTAAT
-CCTACA
-CCTACC
-CCTACG
-CCTACT
-CCTAGA
-CCTAGC
-CCTAGG
-CCTAGT
-CCTATA
-CCTATC
-CCTATG
-CCTATT
-CCTCAA
-CCTCAC
-CCTCAG
-CCTCAT
-CCTCCA
-CCTCCC
-CCTCCG
-CCTCCT
-CCTCGA
-CCTCGC
-CCTCGG
-CCTCGT
-CCTCTA
-CCTCTC
-CCTCTG
-CCTCTT
-CCTGAA
-CCTGAC
-CCTGAG
-CCTGAT
-CCTGCA
-CCTGCC
-CCTGCG
-CCTGCT
-CCTGGA
-CCTGGC
-CCTGGG
-CCTGGT
-CCTGTA
-CCTGTC
-CCTGTG
-CCTGTT
-CCTTAA
-CCTTAC
-CCTTAG
-CCTTAT
-CCTTCA
-CCTTCC
-CCTTCG
-CCTTCT
-CCTTGA
-CCTTGC
-CCTTGG
-CCTTGT
-CCTTTA
-CCTTTC
-CCTTTG
-CCTTTT
-CGAAAA
-CGAAAC
-CGAAAG
-CGAAAT
-CGAACA
-CGAACC
-CGAACG
-CGAACT
-CGAAGA
-CGAAGC
-CGAAGG
-CGAAGT
-CGAATA
-CGAATC
-CGAATG
-CGAATT
-CGACAA
-CGACAC
-CGACAG
-CGACAT
-CGACCA
-CGACCC
-CGACCG
-CGACCT
-CGACGA
-CGACGC
-CGACGG
-CGACGT
-CGACTA
-CGACTC
-CGACTG
-CGACTT
-CGAGAA
-CGAGAC
-CGAGAG
-CGAGAT
-CGAGCA
-CGAGCC
-CGAGCG
-CGAGCT
-CGAGGA
-CGAGGC
-CGAGGG
-CGAGGT
-CGAGTA
-CGAGTC
-CGAGTG
-CGAGTT
-CGATAA
-CGATAC
-CGATAG
-CGATAT
-CGATCA
-CGATCC
-CGATCG
-CGATCT
-CGATGA
-CGATGC
-CGATGG
-CGATGT
-CGATTA
-CGATTC
-CGATTG
-CGATTT
-CGCAAA
-CGCAAC
-CGCAAG
-CGCAAT
-CGCACA
-CGCACC
-CGCACG
-CGCACT
-CGCAGA
-CGCAGC
-CGCAGG
-CGCAGT
-CGCATA
-CGCATC
-CGCATG
-CGCATT
-CGCCAA
-CGCCAC
-CGCCAG
-CGCCAT
-CGCCCA
-CGCCCC
-CGCCCG
-CGCCCT
-CGCCGA
-CGCCGC
-CGCCGG
-CGCCGT
-CGCCTA
-CGCCTC
-CGCCTG
-CGCCTT
-CGCGAA
-CGCGAC
-CGCGAG
-CGCGAT
-CGCGCA
-CGCGCC
-CGCGCG
-CGCGCT
-CGCGGA
-CGCGGC
-CGCGGG
-CGCGGT
-CGCGTA
-CGCGTC
-CGCGTG
-CGCGTT
-CGCTAA
-CGCTAC
-CGCTAG
-CGCTAT
-CGCTCA
-CGCTCC
-CGCTCG
-CGCTCT
-CGCTGA
-CGCTGC
-CGCTGG
-CGCTGT
-CGCTTA
-CGCTTC
-CGCTTG
-CGCTTT
-CGGAAA
-CGGAAC
-CGGAAG
-CGGAAT
-CGGACA
-CGGACC
-CGGACG
-CGGACT
-CGGAGA
-CGGAGC
-CGGAGG
-CGGAGT
-CGGATA
-CGGATC
-CGGATG
-CGGATT
-CGGCAA
-CGGCAC
-CGGCAG
-CGGCAT
-CGGCCA
-CGGCCC
-CGGCCG
-CGGCCT
-CGGCGA
-CGGCGC
-CGGCGG
-CGGCGT
-CGGCTA
-CGGCTC
-CGGCTG
-CGGCTT
-CGGGAA
-CGGGAC
-CGGGAG
-CGGGAT
-CGGGCA
-CGGGCC
-CGGGCG
-CGGGCT
-CGGGGA
-CGGGGC
-CGGGGG
-CGGGGT
-CGGGTA
-CGGGTC
-CGGGTG
-CGGGTT
-CGGTAA
-CGGTAC
-CGGTAG
-CGGTAT
-CGGTCA
-CGGTCC
-CGGTCG
-CGGTCT
-CGGTGA
-CGGTGC
-CGGTGG
-CGGTGT
-CGGTTA
-CGGTTC
-CGGTTG
-CGGTTT
-CGTAAA
-CGTAAC
-CGTAAG
-CGTAAT
-CGTACA
-CGTACC
-CGTACG
-CGTACT
-CGTAGA
-CGTAGC
-CGTAGG
-CGTAGT
-CGTATA
-CGTATC
-CGTATG
-CGTATT
-CGTCAA
-CGTCAC
-CGTCAG
-CGTCAT
-CGTCCA
-CGTCCC
-CGTCCG
-CGTCCT
-CGTCGA
-CGTCGC
-CGTCGG
-CGTCGT
-CGTCTA
-CGTCTC
-CGTCTG
-CGTCTT
-CGTGAA
-CGTGAC
-CGTGAG
-CGTGAT
-CGTGCA
-CGTGCC
-CGTGCG
-CGTGCT
-CGTGGA
-CGTGGC
-CGTGGG
-CGTGGT
-CGTGTA
-CGTGTC
-CGTGTG
-CGTGTT
-CGTTAA
-CGTTAC
-CGTTAG
-CGTTAT
-CGTTCA
-CGTTCC
-CGTTCG
-CGTTCT
-CGTTGA
-CGTTGC
-CGTTGG
-CGTTGT
-CGTTTA
-CGTTTC
-CGTTTG
-CGTTTT
-CTAAAA
-CTAAAC
-CTAAAG
-CTAAAT
-CTAACA
-CTAACC
-CTAACG
-CTAACT
-CTAAGA
-CTAAGC
-CTAAGG
-CTAAGT
-CTAATA
-CTAATC
-CTAATG
-CTAATT
-CTACAA
-CTACAC
-CTACAG
-CTACAT
-CTACCA
-CTACCC
-CTACCG
-CTACCT
-CTACGA
-CTACGC
-CTACGG
-CTACGT
-CTACTA
-CTACTC
-CTACTG
-CTACTT
-CTAGAA
-CTAGAC
-CTAGAG
-CTAGAT
-CTAGCA
-CTAGCC
-CTAGCG
-CTAGCT
-CTAGGA
-CTAGGC
-CTAGGG
-CTAGGT
-CTAGTA
-CTAGTC
-CTAGTG
-CTAGTT
-CTATAA
-CTATAC
-CTATAG
-CTATAT
-CTATCA
-CTATCC
-CTATCG
-CTATCT
-CTATGA
-CTATGC
-CTATGG
-CTATGT
-CTATTA
-CTATTC
-CTATTG
-CTATTT
-CTCAAA
-CTCAAC
-CTCAAG
-CTCAAT
-CTCACA
-CTCACC
-CTCACG
-CTCACT
-CTCAGA
-CTCAGC
-CTCAGG
-CTCAGT
-CTCATA
-CTCATC
-CTCATG
-CTCATT
-CTCCAA
-CTCCAC
-CTCCAG
-CTCCAT
-CTCCCA
-CTCCCC
-CTCCCG
-CTCCCT
-CTCCGA
-CTCCGC
-CTCCGG
-CTCCGT
-CTCCTA
-CTCCTC
-CTCCTG
-CTCCTT
-CTCGAA
-CTCGAC
-CTCGAG
-CTCGAT
-CTCGCA
-CTCGCC
-CTCGCG
-CTCGCT
-CTCGGA
-CTCGGC
-CTCGGG
-CTCGGT
-CTCGTA
-CTCGTC
-CTCGTG
-CTCGTT
-CTCTAA
-CTCTAC
-CTCTAG
-CTCTAT
-CTCTCA
-CTCTCC
-CTCTCG
-CTCTCT
-CTCTGA
-CTCTGC
-CTCTGG
-CTCTGT
-CTCTTA
-CTCTTC
-CTCTTG
-CTCTTT
-CTGAAA
-CTGAAC
-CTGAAG
-CTGAAT
-CTGACA
-CTGACC
-CTGACG
-CTGACT
-CTGAGA
-CTGAGC
-CTGAGG
-CTGAGT
-CTGATA
-CTGATC
-CTGATG
-CTGATT
-CTGCAA
-CTGCAC
-CTGCAG
-CTGCAT
-CTGCCA
-CTGCCC
-CTGCCG
-CTGCCT
-CTGCGA
-CTGCGC
-CTGCGG
-CTGCGT
-CTGCTA
-CTGCTC
-CTGCTG
-CTGCTT
-CTGGAA
-CTGGAC
-CTGGAG
-CTGGAT
-CTGGCA
-CTGGCC
-CTGGCG
-CTGGCT
-CTGGGA
-CTGGGC
-CTGGGG
-CTGGGT
-CTGGTA
-CTGGTC
-CTGGTG
-CTGGTT
-CTGTAA
-CTGTAC
-CTGTAG
-CTGTAT
-CTGTCA
-CTGTCC
-CTGTCG
-CTGTCT
-CTGTGA
-CTGTGC
-CTGTGG
-CTGTGT
-CTGTTA
-CTGTTC
-CTGTTG
-CTGTTT
-CTTAAA
-CTTAAC
-CTTAAG
-CTTAAT
-CTTACA
-CTTACC
-CTTACG
-CTTACT
-CTTAGA
-CTTAGC
-CTTAGG
-CTTAGT
-CTTATA
-CTTATC
-CTTATG
-CTTATT
-CTTCAA
-CTTCAC
-CTTCAG
-CTTCAT
-CTTCCA
-CTTCCC
-CTTCCG
-CTTCCT
-CTTCGA
-CTTCGC
-CTTCGG
-CTTCGT
-CTTCTA
-CTTCTC
-CTTCTG
-CTTCTT
-CTTGAA
-CTTGAC
-CTTGAG
-CTTGAT
-CTTGCA
-CTTGCC
-CTTGCG
-CTTGCT
-CTTGGA
-CTTGGC
-CTTGGG
-CTTGGT
-CTTGTA
-CTTGTC
-CTTGTG
-CTTGTT
-CTTTAA
-CTTTAC
-CTTTAG
-CTTTAT
-CTTTCA
-CTTTCC
-CTTTCG
-CTTTCT
-CTTTGA
-CTTTGC
-CTTTGG
-CTTTGT
-CTTTTA
-CTTTTC
-CTTTTG
-CTTTTT
-GAAAAA
-GAAAAC
-GAAAAG
-GAAAAT
-GAAACA
-GAAACC
-GAAACG
-GAAACT
-GAAAGA
-GAAAGC
-GAAAGG
-GAAAGT
-GAAATA
-GAAATC
-GAAATG
-GAAATT
-GAACAA
-GAACAC
-GAACAG
-GAACAT
-GAACCA
-GAACCC
-GAACCG
-GAACCT
-GAACGA
-GAACGC
-GAACGG
-GAACGT
-GAACTA
-GAACTC
-GAACTG
-GAACTT
-GAAGAA
-GAAGAC
-GAAGAG
-GAAGAT
-GAAGCA
-GAAGCC
-GAAGCG
-GAAGCT
-GAAGGA
-GAAGGC
-GAAGGG
-GAAGGT
-GAAGTA
-GAAGTC
-GAAGTG
-GAAGTT
-GAATAA
-GAATAC
-GAATAG
-GAATAT
-GAATCA
-GAATCC
-GAATCG
-GAATCT
-GAATGA
-GAATGC
-GAATGG
-GAATGT
-GAATTA
-GAATTC
-GAATTG
-GAATTT
-GACAAA
-GACAAC
-GACAAG
-GACAAT
-GACACA
-GACACC
-GACACG
-GACACT
-GACAGA
-GACAGC
-GACAGG
-GACAGT
-GACATA
-GACATC
-GACATG
-GACATT
-GACCAA
-GACCAC
-GACCAG
-GACCAT
-GACCCA
-GACCCC
-GACCCG
-GACCCT
-GACCGA
-GACCGC
-GACCGG
-GACCGT
-GACCTA
-GACCTC
-GACCTG
-GACCTT
-GACGAA
-GACGAC
-GACGAG
-GACGAT
-GACGCA
-GACGCC
-GACGCG
-GACGCT
-GACGGA
-GACGGC
-GACGGG
-GACGGT
-GACGTA
-GACGTC
-GACGTG
-GACGTT
-GACTAA
-GACTAC
-GACTAG
-GACTAT
-GACTCA
-GACTCC
-GACTCG
-GACTCT
-GACTGA
-GACTGC
-GACTGG
-GACTGT
-GACTTA
-GACTTC
-GACTTG
-GACTTT
-GAGAAA
-GAGAAC
-GAGAAG
-GAGAAT
-GAGACA
-GAGACC
-GAGACG
-GAGACT
-GAGAGA
-GAGAGC
-GAGAGG
-GAGAGT
-GAGATA
-GAGATC
-GAGATG
-GAGATT
-GAGCAA
-GAGCAC
-GAGCAG
-GAGCAT
-GAGCCA
-GAGCCC
-GAGCCG
-GAGCCT
-GAGCGA
-GAGCGC
-GAGCGG
-GAGCGT
-GAGCTA
-GAGCTC
-GAGCTG
-GAGCTT
-GAGGAA
-GAGGAC
-GAGGAG
-GAGGAT
-GAGGCA
-GAGGCC
-GAGGCG
-GAGGCT
-GAGGGA
-GAGGGC
-GAGGGG
-GAGGGT
-GAGGTA
-GAGGTC
-GAGGTG
-GAGGTT
-GAGTAA
-GAGTAC
-GAGTAG
-GAGTAT
-GAGTCA
-GAGTCC
-GAGTCG
-GAGTCT
-GAGTGA
-GAGTGC
-GAGTGG
-GAGTGT
-GAGTTA
-GAGTTC
-GAGTTG
-GAGTTT
-GATAAA
-GATAAC
-GATAAG
-GATAAT
-GATACA
-GATACC
-GATACG
-GATACT
-GATAGA
-GATAGC
-GATAGG
-GATAGT
-GATATA
-GATATC
-GATATG
-GATATT
-GATCAA
-GATCAC
-GATCAG
-GATCAT
-GATCCA
-GATCCC
-GATCCG
-GATCCT
-GATCGA
-GATCGC
-GATCGG
-GATCGT
-GATCTA
-GATCTC
-GATCTG
-GATCTT
-GATGAA
-GATGAC
-GATGAG
-GATGAT
-GATGCA
-GATGCC
-GATGCG
-GATGCT
-GATGGA
-GATGGC
-GATGGG
-GATGGT
-GATGTA
-GATGTC
-GATGTG
-GATGTT
-GATTAA
-GATTAC
-GATTAG
-GATTAT
-GATTCA
-GATTCC
-GATTCG
-GATTCT
-GATTGA
-GATTGC
-GATTGG
-GATTGT
-GATTTA
-GATTTC
-GATTTG
-GATTTT
-GCAAAA
-GCAAAC
-GCAAAG
-GCAAAT
-GCAACA
-GCAACC
-GCAACG
-GCAACT
-GCAAGA
-GCAAGC
-GCAAGG
-GCAAGT
-GCAATA
-GCAATC
-GCAATG
-GCAATT
-GCACAA
-GCACAC
-GCACAG
-GCACAT
-GCACCA
-GCACCC
-GCACCG
-GCACCT
-GCACGA
-GCACGC
-GCACGG
-GCACGT
-GCACTA
-GCACTC
-GCACTG
-GCACTT
-GCAGAA
-GCAGAC
-GCAGAG
-GCAGAT
-GCAGCA
-GCAGCC
-GCAGCG
-GCAGCT
-GCAGGA
-GCAGGC
-GCAGGG
-GCAGGT
-GCAGTA
-GCAGTC
-GCAGTG
-GCAGTT
-GCATAA
-GCATAC
-GCATAG
-GCATAT
-GCATCA
-GCATCC
-GCATCG
-GCATCT
-GCATGA
-GCATGC
-GCATGG
-GCATGT
-GCATTA
-GCATTC
-GCATTG
-GCATTT
-GCCAAA
-GCCAAC
-GCCAAG
-GCCAAT
-GCCACA
-GCCACC
-GCCACG
-GCCACT
-GCCAGA
-GCCAGC
-GCCAGG
-GCCAGT
-GCCATA
-GCCATC
-GCCATG
-GCCATT
-GCCCAA
-GCCCAC
-GCCCAG
-GCCCAT
-GCCCCA
-GCCCCC
-GCCCCG
-GCCCCT
-GCCCGA
-GCCCGC
-GCCCGG
-GCCCGT
-GCCCTA
-GCCCTC
-GCCCTG
-GCCCTT
-GCCGAA
-GCCGAC
-GCCGAG
-GCCGAT
-GCCGCA
-GCCGCC
-GCCGCG
-GCCGCT
-GCCGGA
-GCCGGC
-GCCGGG
-GCCGGT
-GCCGTA
-GCCGTC
-GCCGTG
-GCCGTT
-GCCTAA
-GCCTAC
-GCCTAG
-GCCTAT
-GCCTCA
-GCCTCC
-GCCTCG
-GCCTCT
-GCCTGA
-GCCTGC
-GCCTGG
-GCCTGT
-GCCTTA
-GCCTTC
-GCCTTG
-GCCTTT
-GCGAAA
-GCGAAC
-GCGAAG
-GCGAAT
-GCGACA
-GCGACC
-GCGACG
-GCGACT
-GCGAGA
-GCGAGC
-GCGAGG
-GCGAGT
-GCGATA
-GCGATC
-GCGATG
-GCGATT
-GCGCAA
-GCGCAC
-GCGCAG
-GCGCAT
-GCGCCA
-GCGCCC
-GCGCCG
-GCGCCT
-GCGCGA
-GCGCGC
-GCGCGG
-GCGCGT
-GCGCTA
-GCGCTC
-GCGCTG
-GCGCTT
-GCGGAA
-GCGGAC
-GCGGAG
-GCGGAT
-GCGGCA
-GCGGCC
-GCGGCG
-GCGGCT
-GCGGGA
-GCGGGC
-GCGGGG
-GCGGGT
-GCGGTA
-GCGGTC
-GCGGTG
-GCGGTT
-GCGTAA
-GCGTAC
-GCGTAG
-GCGTAT
-GCGTCA
-GCGTCC
-GCGTCG
-GCGTCT
-GCGTGA
-GCGTGC
-GCGTGG
-GCGTGT
-GCGTTA
-GCGTTC
-GCGTTG
-GCGTTT
-GCTAAA
-GCTAAC
-GCTAAG
-GCTAAT
-GCTACA
-GCTACC
-GCTACG
-GCTACT
-GCTAGA
-GCTAGC
-GCTAGG
-GCTAGT
-GCTATA
-GCTATC
-GCTATG
-GCTATT
-GCTCAA
-GCTCAC
-GCTCAG
-GCTCAT
-GCTCCA
-GCTCCC
-GCTCCG
-GCTCCT
-GCTCGA
-GCTCGC
-GCTCGG
-GCTCGT
-GCTCTA
-GCTCTC
-GCTCTG
-GCTCTT
-GCTGAA
-GCTGAC
-GCTGAG
-GCTGAT
-GCTGCA
-GCTGCC
-GCTGCG
-GCTGCT
-GCTGGA
-GCTGGC
-GCTGGG
-GCTGGT
-GCTGTA
-GCTGTC
-GCTGTG
-GCTGTT
-GCTTAA
-GCTTAC
-GCTTAG
-GCTTAT
-GCTTCA
-GCTTCC
-GCTTCG
-GCTTCT
-GCTTGA
-GCTTGC
-GCTTGG
-GCTTGT
-GCTTTA
-GCTTTC
-GCTTTG
-GCTTTT
-GGAAAA
-GGAAAC
-GGAAAG
-GGAAAT
-GGAACA
-GGAACC
-GGAACG
-GGAACT
-GGAAGA
-GGAAGC
-GGAAGG
-GGAAGT
-GGAATA
-GGAATC
-GGAATG
-GGAATT
-GGACAA
-GGACAC
-GGACAG
-GGACAT
-GGACCA
-GGACCC
-GGACCG
-GGACCT
-GGACGA
-GGACGC
-GGACGG
-GGACGT
-GGACTA
-GGACTC
-GGACTG
-GGACTT
-GGAGAA
-GGAGAC
-GGAGAG
-GGAGAT
-GGAGCA
-GGAGCC
-GGAGCG
-GGAGCT
-GGAGGA
-GGAGGC
-GGAGGG
-GGAGGT
-GGAGTA
-GGAGTC
-GGAGTG
-GGAGTT
-GGATAA
-GGATAC
-GGATAG
-GGATAT
-GGATCA
-GGATCC
-GGATCG
-GGATCT
-GGATGA
-GGATGC
-GGATGG
-GGATGT
-GGATTA
-GGATTC
-GGATTG
-GGATTT
-GGCAAA
-GGCAAC
-GGCAAG
-GGCAAT
-GGCACA
-GGCACC
-GGCACG
-GGCACT
-GGCAGA
-GGCAGC
-GGCAGG
-GGCAGT
-GGCATA
-GGCATC
-GGCATG
-GGCATT
-GGCCAA
-GGCCAC
-GGCCAG
-GGCCAT
-GGCCCA
-GGCCCC
-GGCCCG
-GGCCCT
-GGCCGA
-GGCCGC
-GGCCGG
-GGCCGT
-GGCCTA
-GGCCTC
-GGCCTG
-GGCCTT
-GGCGAA
-GGCGAC
-GGCGAG
-GGCGAT
-GGCGCA
-GGCGCC
-GGCGCG
-GGCGCT
-GGCGGA
-GGCGGC
-GGCGGG
-GGCGGT
-GGCGTA
-GGCGTC
-GGCGTG
-GGCGTT
-GGCTAA
-GGCTAC
-GGCTAG
-GGCTAT
-GGCTCA
-GGCTCC
-GGCTCG
-GGCTCT
-GGCTGA
-GGCTGC
-GGCTGG
-GGCTGT
-GGCTTA
-GGCTTC
-GGCTTG
-GGCTTT
-GGGAAA
-GGGAAC
-GGGAAG
-GGGAAT
-GGGACA
-GGGACC
-GGGACG
-GGGACT
-GGGAGA
-GGGAGC
-GGGAGG
-GGGAGT
-GGGATA
-GGGATC
-GGGATG
-GGGATT
-GGGCAA
-GGGCAC
-GGGCAG
-GGGCAT
-GGGCCA
-GGGCCC
-GGGCCG
-GGGCCT
-GGGCGA
-GGGCGC
-GGGCGG
-GGGCGT
-GGGCTA
-GGGCTC
-GGGCTG
-GGGCTT
-GGGGAA
-GGGGAC
-GGGGAG
-GGGGAT
-GGGGCA
-GGGGCC
-GGGGCG
-GGGGCT
-GGGGGA
-GGGGGC
-GGGGGG
-GGGGGT
-GGGGTA
-GGGGTC
-GGGGTG
-GGGGTT
-GGGTAA
-GGGTAC
-GGGTAG
-GGGTAT
-GGGTCA
-GGGTCC
-GGGTCG
-GGGTCT
-GGGTGA
-GGGTGC
-GGGTGG
-GGGTGT
-GGGTTA
-GGGTTC
-GGGTTG
-GGGTTT
-GGTAAA
-GGTAAC
-GGTAAG
-GGTAAT
-GGTACA
-GGTACC
-GGTACG
-GGTACT
-GGTAGA
-GGTAGC
-GGTAGG
-GGTAGT
-GGTATA
-GGTATC
-GGTATG
-GGTATT
-GGTCAA
-GGTCAC
-GGTCAG
-GGTCAT
-GGTCCA
-GGTCCC
-GGTCCG
-GGTCCT
-GGTCGA
-GGTCGC
-GGTCGG
-GGTCGT
-GGTCTA
-GGTCTC
-GGTCTG
-GGTCTT
-GGTGAA
-GGTGAC
-GGTGAG
-GGTGAT
-GGTGCA
-GGTGCC
-GGTGCG
-GGTGCT
-GGTGGA
-GGTGGC
-GGTGGG
-GGTGGT
-GGTGTA
-GGTGTC
-GGTGTG
-GGTGTT
-GGTTAA
-GGTTAC
-GGTTAG
-GGTTAT
-GGTTCA
-GGTTCC
-GGTTCG
-GGTTCT
-GGTTGA
-GGTTGC
-GGTTGG
-GGTTGT
-GGTTTA
-GGTTTC
-GGTTTG
-GGTTTT
-GTAAAA
-GTAAAC
-GTAAAG
-GTAAAT
-GTAACA
-GTAACC
-GTAACG
-GTAACT
-GTAAGA
-GTAAGC
-GTAAGG
-GTAAGT
-GTAATA
-GTAATC
-GTAATG
-GTAATT
-GTACAA
-GTACAC
-GTACAG
-GTACAT
-GTACCA
-GTACCC
-GTACCG
-GTACCT
-GTACGA
-GTACGC
-GTACGG
-GTACGT
-GTACTA
-GTACTC
-GTACTG
-GTACTT
-GTAGAA
-GTAGAC
-GTAGAG
-GTAGAT
-GTAGCA
-GTAGCC
-GTAGCG
-GTAGCT
-GTAGGA
-GTAGGC
-GTAGGG
-GTAGGT
-GTAGTA
-GTAGTC
-GTAGTG
-GTAGTT
-GTATAA
-GTATAC
-GTATAG
-GTATAT
-GTATCA
-GTATCC
-GTATCG
-GTATCT
-GTATGA
-GTATGC
-GTATGG
-GTATGT
-GTATTA
-GTATTC
-GTATTG
-GTATTT
-GTCAAA
-GTCAAC
-GTCAAG
-GTCAAT
-GTCACA
-GTCACC
-GTCACG
-GTCACT
-GTCAGA
-GTCAGC
-GTCAGG
-GTCAGT
-GTCATA
-GTCATC
-GTCATG
-GTCATT
-GTCCAA
-GTCCAC
-GTCCAG
-GTCCAT
-GTCCCA
-GTCCCC
-GTCCCG
-GTCCCT
-GTCCGA
-GTCCGC
-GTCCGG
-GTCCGT
-GTCCTA
-GTCCTC
-GTCCTG
-GTCCTT
-GTCGAA
-GTCGAC
-GTCGAG
-GTCGAT
-GTCGCA
-GTCGCC
-GTCGCG
-GTCGCT
-GTCGGA
-GTCGGC
-GTCGGG
-GTCGGT
-GTCGTA
-GTCGTC
-GTCGTG
-GTCGTT
-GTCTAA
-GTCTAC
-GTCTAG
-GTCTAT
-GTCTCA
-GTCTCC
-GTCTCG
-GTCTCT
-GTCTGA
-GTCTGC
-GTCTGG
-GTCTGT
-GTCTTA
-GTCTTC
-GTCTTG
-GTCTTT
-GTGAAA
-GTGAAC
-GTGAAG
-GTGAAT
-GTGACA
-GTGACC
-GTGACG
-GTGACT
-GTGAGA
-GTGAGC
-GTGAGG
-GTGAGT
-GTGATA
-GTGATC
-GTGATG
-GTGATT
-GTGCAA
-GTGCAC
-GTGCAG
-GTGCAT
-GTGCCA
-GTGCCC
-GTGCCG
-GTGCCT
-GTGCGA
-GTGCGC
-GTGCGG
-GTGCGT
-GTGCTA
-GTGCTC
-GTGCTG
-GTGCTT
-GTGGAA
-GTGGAC
-GTGGAG
-GTGGAT
-GTGGCA
-GTGGCC
-GTGGCG
-GTGGCT
-GTGGGA
-GTGGGC
-GTGGGG
-GTGGGT
-GTGGTA
-GTGGTC
-GTGGTG
-GTGGTT
-GTGTAA
-GTGTAC
-GTGTAG
-GTGTAT
-GTGTCA
-GTGTCC
-GTGTCG
-GTGTCT
-GTGTGA
-GTGTGC
-GTGTGG
-GTGTGT
-GTGTTA
-GTGTTC
-GTGTTG
-GTGTTT
-GTTAAA
-GTTAAC
-GTTAAG
-GTTAAT
-GTTACA
-GTTACC
-GTTACG
-GTTACT
-GTTAGA
-GTTAGC
-GTTAGG
-GTTAGT
-GTTATA
-GTTATC
-GTTATG
-GTTATT
-GTTCAA
-GTTCAC
-GTTCAG
-GTTCAT
-GTTCCA
-GTTCCC
-GTTCCG
-GTTCCT
-GTTCGA
-GTTCGC
-GTTCGG
-GTTCGT
-GTTCTA
-GTTCTC
-GTTCTG
-GTTCTT
-GTTGAA
-GTTGAC
-GTTGAG
-GTTGAT
-GTTGCA
-GTTGCC
-GTTGCG
-GTTGCT
-GTTGGA
-GTTGGC
-GTTGGG
-GTTGGT
-GTTGTA
-GTTGTC
-GTTGTG
-GTTGTT
-GTTTAA
-GTTTAC
-GTTTAG
-GTTTAT
-GTTTCA
-GTTTCC
-GTTTCG
-GTTTCT
-GTTTGA
-GTTTGC
-GTTTGG
-GTTTGT
-GTTTTA
-GTTTTC
-GTTTTG
-GTTTTT
-TAAAAA
-TAAAAC
-TAAAAG
-TAAAAT
-TAAACA
-TAAACC
-TAAACG
-TAAACT
-TAAAGA
-TAAAGC
-TAAAGG
-TAAAGT
-TAAATA
-TAAATC
-TAAATG
-TAAATT
-TAACAA
-TAACAC
-TAACAG
-TAACAT
-TAACCA
-TAACCC
-TAACCG
-TAACCT
-TAACGA
-TAACGC
-TAACGG
-TAACGT
-TAACTA
-TAACTC
-TAACTG
-TAACTT
-TAAGAA
-TAAGAC
-TAAGAG
-TAAGAT
-TAAGCA
-TAAGCC
-TAAGCG
-TAAGCT
-TAAGGA
-TAAGGC
-TAAGGG
-TAAGGT
-TAAGTA
-TAAGTC
-TAAGTG
-TAAGTT
-TAATAA
-TAATAC
-TAATAG
-TAATAT
-TAATCA
-TAATCC
-TAATCG
-TAATCT
-TAATGA
-TAATGC
-TAATGG
-TAATGT
-TAATTA
-TAATTC
-TAATTG
-TAATTT
-TACAAA
-TACAAC
-TACAAG
-TACAAT
-TACACA
-TACACC
-TACACG
-TACACT
-TACAGA
-TACAGC
-TACAGG
-TACAGT
-TACATA
-TACATC
-TACATG
-TACATT
-TACCAA
-TACCAC
-TACCAG
-TACCAT
-TACCCA
-TACCCC
-TACCCG
-TACCCT
-TACCGA
-TACCGC
-TACCGG
-TACCGT
-TACCTA
-TACCTC
-TACCTG
-TACCTT
-TACGAA
-TACGAC
-TACGAG
-TACGAT
-TACGCA
-TACGCC
-TACGCG
-TACGCT
-TACGGA
-TACGGC
-TACGGG
-TACGGT
-TACGTA
-TACGTC
-TACGTG
-TACGTT
-TACTAA
-TACTAC
-TACTAG
-TACTAT
-TACTCA
-TACTCC
-TACTCG
-TACTCT
-TACTGA
-TACTGC
-TACTGG
-TACTGT
-TACTTA
-TACTTC
-TACTTG
-TACTTT
-TAGAAA
-TAGAAC
-TAGAAG
-TAGAAT
-TAGACA
-TAGACC
-TAGACG
-TAGACT
-TAGAGA
-TAGAGC
-TAGAGG
-TAGAGT
-TAGATA
-TAGATC
-TAGATG
-TAGATT
-TAGCAA
-TAGCAC
-TAGCAG
-TAGCAT
-TAGCCA
-TAGCCC
-TAGCCG
-TAGCCT
-TAGCGA
-TAGCGC
-TAGCGG
-TAGCGT
-TAGCTA
-TAGCTC
-TAGCTG
-TAGCTT
-TAGGAA
-TAGGAC
-TAGGAG
-TAGGAT
-TAGGCA
-TAGGCC
-TAGGCG
-TAGGCT
-TAGGGA
-TAGGGC
-TAGGGG
-TAGGGT
-TAGGTA
-TAGGTC
-TAGGTG
-TAGGTT
-TAGTAA
-TAGTAC
-TAGTAG
-TAGTAT
-TAGTCA
-TAGTCC
-TAGTCG
-TAGTCT
-TAGTGA
-TAGTGC
-TAGTGG
-TAGTGT
-TAGTTA
-TAGTTC
-TAGTTG
-TAGTTT
-TATAAA
-TATAAC
-TATAAG
-TATAAT
-TATACA
-TATACC
-TATACG
-TATACT
-TATAGA
-TATAGC
-TATAGG
-TATAGT
-TATATA
-TATATC
-TATATG
-TATATT
-TATCAA
-TATCAC
-TATCAG
-TATCAT
-TATCCA
-TATCCC
-TATCCG
-TATCCT
-TATCGA
-TATCGC
-TATCGG
-TATCGT
-TATCTA
-TATCTC
-TATCTG
-TATCTT
-TATGAA
-TATGAC
-TATGAG
-TATGAT
-TATGCA
-TATGCC
-TATGCG
-TATGCT
-TATGGA
-TATGGC
-TATGGG
-TATGGT
-TATGTA
-TATGTC
-TATGTG
-TATGTT
-TATTAA
-TATTAC
-TATTAG
-TATTAT
-TATTCA
-TATTCC
-TATTCG
-TATTCT
-TATTGA
-TATTGC
-TATTGG
-TATTGT
-TATTTA
-TATTTC
-TATTTG
-TATTTT
-TCAAAA
-TCAAAC
-TCAAAG
-TCAAAT
-TCAACA
-TCAACC
-TCAACG
-TCAACT
-TCAAGA
-TCAAGC
-TCAAGG
-TCAAGT
-TCAATA
-TCAATC
-TCAATG
-TCAATT
-TCACAA
-TCACAC
-TCACAG
-TCACAT
-TCACCA
-TCACCC
-TCACCG
-TCACCT
-TCACGA
-TCACGC
-TCACGG
-TCACGT
-TCACTA
-TCACTC
-TCACTG
-TCACTT
-TCAGAA
-TCAGAC
-TCAGAG
-TCAGAT
-TCAGCA
-TCAGCC
-TCAGCG
-TCAGCT
-TCAGGA
-TCAGGC
-TCAGGG
-TCAGGT
-TCAGTA
-TCAGTC
-TCAGTG
-TCAGTT
-TCATAA
-TCATAC
-TCATAG
-TCATAT
-TCATCA
-TCATCC
-TCATCG
-TCATCT
-TCATGA
-TCATGC
-TCATGG
-TCATGT
-TCATTA
-TCATTC
-TCATTG
-TCATTT
-TCCAAA
-TCCAAC
-TCCAAG
-TCCAAT
-TCCACA
-TCCACC
-TCCACG
-TCCACT
-TCCAGA
-TCCAGC
-TCCAGG
-TCCAGT
-TCCATA
-TCCATC
-TCCATG
-TCCATT
-TCCCAA
-TCCCAC
-TCCCAG
-TCCCAT
-TCCCCA
-TCCCCC
-TCCCCG
-TCCCCT
-TCCCGA
-TCCCGC
-TCCCGG
-TCCCGT
-TCCCTA
-TCCCTC
-TCCCTG
-TCCCTT
-TCCGAA
-TCCGAC
-TCCGAG
-TCCGAT
-TCCGCA
-TCCGCC
-TCCGCG
-TCCGCT
-TCCGGA
-TCCGGC
-TCCGGG
-TCCGGT
-TCCGTA
-TCCGTC
-TCCGTG
-TCCGTT
-TCCTAA
-TCCTAC
-TCCTAG
-TCCTAT
-TCCTCA
-TCCTCC
-TCCTCG
-TCCTCT
-TCCTGA
-TCCTGC
-TCCTGG
-TCCTGT
-TCCTTA
-TCCTTC
-TCCTTG
-TCCTTT
-TCGAAA
-TCGAAC
-TCGAAG
-TCGAAT
-TCGACA
-TCGACC
-TCGACG
-TCGACT
-TCGAGA
-TCGAGC
-TCGAGG
-TCGAGT
-TCGATA
-TCGATC
-TCGATG
-TCGATT
-TCGCAA
-TCGCAC
-TCGCAG
-TCGCAT
-TCGCCA
-TCGCCC
-TCGCCG
-TCGCCT
-TCGCGA
-TCGCGC
-TCGCGG
-TCGCGT
-TCGCTA
-TCGCTC
-TCGCTG
-TCGCTT
-TCGGAA
-TCGGAC
-TCGGAG
-TCGGAT
-TCGGCA
-TCGGCC
-TCGGCG
-TCGGCT
-TCGGGA
-TCGGGC
-TCGGGG
-TCGGGT
-TCGGTA
-TCGGTC
-TCGGTG
-TCGGTT
-TCGTAA
-TCGTAC
-TCGTAG
-TCGTAT
-TCGTCA
-TCGTCC
-TCGTCG
-TCGTCT
-TCGTGA
-TCGTGC
-TCGTGG
-TCGTGT
-TCGTTA
-TCGTTC
-TCGTTG
-TCGTTT
-TCTAAA
-TCTAAC
-TCTAAG
-TCTAAT
-TCTACA
-TCTACC
-TCTACG
-TCTACT
-TCTAGA
-TCTAGC
-TCTAGG
-TCTAGT
-TCTATA
-TCTATC
-TCTATG
-TCTATT
-TCTCAA
-TCTCAC
-TCTCAG
-TCTCAT
-TCTCCA
-TCTCCC
-TCTCCG
-TCTCCT
-TCTCGA
-TCTCGC
-TCTCGG
-TCTCGT
-TCTCTA
-TCTCTC
-TCTCTG
-TCTCTT
-TCTGAA
-TCTGAC
-TCTGAG
-TCTGAT
-TCTGCA
-TCTGCC
-TCTGCG
-TCTGCT
-TCTGGA
-TCTGGC
-TCTGGG
-TCTGGT
-TCTGTA
-TCTGTC
-TCTGTG
-TCTGTT
-TCTTAA
-TCTTAC
-TCTTAG
-TCTTAT
-TCTTCA
-TCTTCC
-TCTTCG
-TCTTCT
-TCTTGA
-TCTTGC
-TCTTGG
-TCTTGT
-TCTTTA
-TCTTTC
-TCTTTG
-TCTTTT
-TGAAAA
-TGAAAC
-TGAAAG
-TGAAAT
-TGAACA
-TGAACC
-TGAACG
-TGAACT
-TGAAGA
-TGAAGC
-TGAAGG
-TGAAGT
-TGAATA
-TGAATC
-TGAATG
-TGAATT
-TGACAA
-TGACAC
-TGACAG
-TGACAT
-TGACCA
-TGACCC
-TGACCG
-TGACCT
-TGACGA
-TGACGC
-TGACGG
-TGACGT
-TGACTA
-TGACTC
-TGACTG
-TGACTT
-TGAGAA
-TGAGAC
-TGAGAG
-TGAGAT
-TGAGCA
-TGAGCC
-TGAGCG
-TGAGCT
-TGAGGA
-TGAGGC
-TGAGGG
-TGAGGT
-TGAGTA
-TGAGTC
-TGAGTG
-TGAGTT
-TGATAA
-TGATAC
-TGATAG
-TGATAT
-TGATCA
-TGATCC
-TGATCG
-TGATCT
-TGATGA
-TGATGC
-TGATGG
-TGATGT
-TGATTA
-TGATTC
-TGATTG
-TGATTT
-TGCAAA
-TGCAAC
-TGCAAG
-TGCAAT
-TGCACA
-TGCACC
-TGCACG
-TGCACT
-TGCAGA
-TGCAGC
-TGCAGG
-TGCAGT
-TGCATA
-TGCATC
-TGCATG
-TGCATT
-TGCCAA
-TGCCAC
-TGCCAG
-TGCCAT
-TGCCCA
-TGCCCC
-TGCCCG
-TGCCCT
-TGCCGA
-TGCCGC
-TGCCGG
-TGCCGT
-TGCCTA
-TGCCTC
-TGCCTG
-TGCCTT
-TGCGAA
-TGCGAC
-TGCGAG
-TGCGAT
-TGCGCA
-TGCGCC
-TGCGCG
-TGCGCT
-TGCGGA
-TGCGGC
-TGCGGG
-TGCGGT
-TGCGTA
-TGCGTC
-TGCGTG
-TGCGTT
-TGCTAA
-TGCTAC
-TGCTAG
-TGCTAT
-TGCTCA
-TGCTCC
-TGCTCG
-TGCTCT
-TGCTGA
-TGCTGC
-TGCTGG
-TGCTGT
-TGCTTA
-TGCTTC
-TGCTTG
-TGCTTT
-TGGAAA
-TGGAAC
-TGGAAG
-TGGAAT
-TGGACA
-TGGACC
-TGGACG
-TGGACT
-TGGAGA
-TGGAGC
-TGGAGG
-TGGAGT
-TGGATA
-TGGATC
-TGGATG
-TGGATT
-TGGCAA
-TGGCAC
-TGGCAG
-TGGCAT
-TGGCCA
-TGGCCC
-TGGCCG
-TGGCCT
-TGGCGA
-TGGCGC
-TGGCGG
-TGGCGT
-TGGCTA
-TGGCTC
-TGGCTG
-TGGCTT
-TGGGAA
-TGGGAC
-TGGGAG
-TGGGAT
-TGGGCA
-TGGGCC
-TGGGCG
-TGGGCT
-TGGGGA
-TGGGGC
-TGGGGG
-TGGGGT
-TGGGTA
-TGGGTC
-TGGGTG
-TGGGTT
-TGGTAA
-TGGTAC
-TGGTAG
-TGGTAT
-TGGTCA
-TGGTCC
-TGGTCG
-TGGTCT
-TGGTGA
-TGGTGC
-TGGTGG
-TGGTGT
-TGGTTA
-TGGTTC
-TGGTTG
-TGGTTT
-TGTAAA
-TGTAAC
-TGTAAG
-TGTAAT
-TGTACA
-TGTACC
-TGTACG
-TGTACT
-TGTAGA
-TGTAGC
-TGTAGG
-TGTAGT
-TGTATA
-TGTATC
-TGTATG
-TGTATT
-TGTCAA
-TGTCAC
-TGTCAG
-TGTCAT
-TGTCCA
-TGTCCC
-TGTCCG
-TGTCCT
-TGTCGA
-TGTCGC
-TGTCGG
-TGTCGT
-TGTCTA
-TGTCTC
-TGTCTG
-TGTCTT
-TGTGAA
-TGTGAC
-TGTGAG
-TGTGAT
-TGTGCA
-TGTGCC
-TGTGCG
-TGTGCT
-TGTGGA
-TGTGGC
-TGTGGG
-TGTGGT
-TGTGTA
-TGTGTC
-TGTGTG
-TGTGTT
-TGTTAA
-TGTTAC
-TGTTAG
-TGTTAT
-TGTTCA
-TGTTCC
-TGTTCG
-TGTTCT
-TGTTGA
-TGTTGC
-TGTTGG
-TGTTGT
-TGTTTA
-TGTTTC
-TGTTTG
-TGTTTT
-TTAAAA
-TTAAAC
-TTAAAG
-TTAAAT
-TTAACA
-TTAACC
-TTAACG
-TTAACT
-TTAAGA
-TTAAGC
-TTAAGG
-TTAAGT
-TTAATA
-TTAATC
-TTAATG
-TTAATT
-TTACAA
-TTACAC
-TTACAG
-TTACAT
-TTACCA
-TTACCC
-TTACCG
-TTACCT
-TTACGA
-TTACGC
-TTACGG
-TTACGT
-TTACTA
-TTACTC
-TTACTG
-TTACTT
-TTAGAA
-TTAGAC
-TTAGAG
-TTAGAT
-TTAGCA
-TTAGCC
-TTAGCG
-TTAGCT
-TTAGGA
-TTAGGC
-TTAGGG
-TTAGGT
-TTAGTA
-TTAGTC
-TTAGTG
-TTAGTT
-TTATAA
-TTATAC
-TTATAG
-TTATAT
-TTATCA
-TTATCC
-TTATCG
-TTATCT
-TTATGA
-TTATGC
-TTATGG
-TTATGT
-TTATTA
-TTATTC
-TTATTG
-TTATTT
-TTCAAA
-TTCAAC
-TTCAAG
-TTCAAT
-TTCACA
-TTCACC
-TTCACG
-TTCACT
-TTCAGA
-TTCAGC
-TTCAGG
-TTCAGT
-TTCATA
-TTCATC
-TTCATG
-TTCATT
-TTCCAA
-TTCCAC
-TTCCAG
-TTCCAT
-TTCCCA
-TTCCCC
-TTCCCG
-TTCCCT
-TTCCGA
-TTCCGC
-TTCCGG
-TTCCGT
-TTCCTA
-TTCCTC
-TTCCTG
-TTCCTT
-TTCGAA
-TTCGAC
-TTCGAG
-TTCGAT
-TTCGCA
-TTCGCC
-TTCGCG
-TTCGCT
-TTCGGA
-TTCGGC
-TTCGGG
-TTCGGT
-TTCGTA
-TTCGTC
-TTCGTG
-TTCGTT
-TTCTAA
-TTCTAC
-TTCTAG
-TTCTAT
-TTCTCA
-TTCTCC
-TTCTCG
-TTCTCT
-TTCTGA
-TTCTGC
-TTCTGG
-TTCTGT
-TTCTTA
-TTCTTC
-TTCTTG
-TTCTTT
-TTGAAA
-TTGAAC
-TTGAAG
-TTGAAT
-TTGACA
-TTGACC
-TTGACG
-TTGACT
-TTGAGA
-TTGAGC
-TTGAGG
-TTGAGT
-TTGATA
-TTGATC
-TTGATG
-TTGATT
-TTGCAA
-TTGCAC
-TTGCAG
-TTGCAT
-TTGCCA
-TTGCCC
-TTGCCG
-TTGCCT
-TTGCGA
-TTGCGC
-TTGCGG
-TTGCGT
-TTGCTA
-TTGCTC
-TTGCTG
-TTGCTT
-TTGGAA
-TTGGAC
-TTGGAG
-TTGGAT
-TTGGCA
-TTGGCC
-TTGGCG
-TTGGCT
-TTGGGA
-TTGGGC
-TTGGGG
-TTGGGT
-TTGGTA
-TTGGTC
-TTGGTG
-TTGGTT
-TTGTAA
-TTGTAC
-TTGTAG
-TTGTAT
-TTGTCA
-TTGTCC
-TTGTCG
-TTGTCT
-TTGTGA
-TTGTGC
-TTGTGG
-TTGTGT
-TTGTTA
-TTGTTC
-TTGTTG
-TTGTTT
-TTTAAA
-TTTAAC
-TTTAAG
-TTTAAT
-TTTACA
-TTTACC
-TTTACG
-TTTACT
-TTTAGA
-TTTAGC
-TTTAGG
-TTTAGT
-TTTATA
-TTTATC
-TTTATG
-TTTATT
-TTTCAA
-TTTCAC
-TTTCAG
-TTTCAT
-TTTCCA
-TTTCCC
-TTTCCG
-TTTCCT
-TTTCGA
-TTTCGC
-TTTCGG
-TTTCGT
-TTTCTA
-TTTCTC
-TTTCTG
-TTTCTT
-TTTGAA
-TTTGAC
-TTTGAG
-TTTGAT
-TTTGCA
-TTTGCC
-TTTGCG
-TTTGCT
-TTTGGA
-TTTGGC
-TTTGGG
-TTTGGT
-TTTGTA
-TTTGTC
-TTTGTG
-TTTGTT
-TTTTAA
-TTTTAC
-TTTTAG
-TTTTAT
-TTTTCA
-TTTTCC
-TTTTCG
-TTTTCT
-TTTTGA
-TTTTGC
-TTTTGG
-TTTTGT
-TTTTTA
-TTTTTC
-TTTTTG
-TTTTTT

general_utils.py

@@ -1,304 +0,0 @@
-# coding=utf-8
-
-import pandas as pd
-import os
-import numpy as np
-import subprocess
-import shutil
-""" Library for general utils, such as dataframe properties checking,
-creating directories, checking files, etc.
-"""
-
-
-def check_expected_columns(df: pd.DataFrame, expected_columns: list) -> bool:
-    """
-    Checks if a DataFrame contains the expected columns.
-    
-    Parameters
-    ----------
-    df : pd.DataFrame
-        The input DataFrame to be checked.
-    expected_columns : list
-        A list of columns that are expected to be present in the DataFrame.
-        
-    Returns
-    -------
-    bool
-        True if all expected columns are present in the DataFrame, False otherwise.
-        
-    Raises
-    ------
-    ValueError
-        If any of the expected columns are not present in the DataFrame.
-        
-    Examples
-    --------
-    >>> df = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})
-    >>> check_expected_columns(df, ['A', 'B'])
-    True
-    
-    >>> check_expected_columns(df, ['A', 'C'])
-    ValueError: The following columns are missing: ['C']
-    """
-    
-    missing_columns = [col for col in expected_columns if col not in df.columns]
-    
-    if missing_columns:
-        raise ValueError(f"The following columns are missing: {missing_columns}")
-    
-    return True
-
-
-def is_valid_primary_key(df: pd.DataFrame, column_name: str) -> bool:
-    """
-    Checks if a specified column in a DataFrame can serve as a valid primary key.
-    
-    Parameters
-    ----------
-    df : pd.DataFrame
-        The input DataFrame to be checked.
-    column_name : str
-        The name of the column to check.
-        
-    Returns
-    -------
-    bool
-        True if the column can serve as a valid primary key, False otherwise.
-        
-    Raises
-    ------
-    ValueError
-        If the specified column does not exist in the DataFrame.
-        
-    Examples
-    --------
-    >>> df = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})
-    >>> is_valid_primary_key(df, 'A')
-    True
-    
-    >>> df = pd.DataFrame({'A': [1, 2, 2], 'B': [4, 5, 6]})
-    >>> is_valid_primary_key(df, 'A')
-    False
-    """
-    
-    if column_name not in df.columns:
-        raise ValueError(f"Column '{column_name}' does not exist in the DataFrame.")
-    
-    # Check for NaN values
-    if df[column_name].isnull().any():
-        return False
-    
-    # Check for unique values
-    if not df[column_name].is_unique:
-        return False
-    
-    return True
-
-def get_non_empty_files(start_path: str, extensions: tuple = ('.fasta', '.fna')) -> str:
-    """
-    Generator that yields non-empty files from a specified directory and its subdirectories based on the given extensions.
-
-    :param start_path: The path to the directory from which to start the search.
-    :type start_path: str
-
-    :param extensions: A tuple of file extensions to look for (default is ('.fasta', '.fna')).
-                       The function also automatically checks for compressed versions with '.gz'.
-    :type extensions: tuple
-
-    :return: Yields filenames that match the specified extensions and are non-empty.
-    :rtype: str
-    """
-    
-    for dirpath, _, filenames in os.walk(start_path):
-        for filename in filenames:
-            filepath = os.path.join(dirpath, filename)
-            if any(filename.endswith(ext) or filename.endswith(ext + '.gz') for ext in extensions) and os.path.getsize(filepath) > 0:
-                yield filename
-
-
-
-def truncate_zero_columns(arr: np.ndarray) -> np.ndarray:
-    """
-    Truncate all trailing columns composed entirely of zeros in a given 2D numpy array.
-    
-    :param arr: Input 2D numpy array.
-    :type arr: np.ndarray
-
-    :return: A new array with trailing zero columns removed.
-    :rtype: np.ndarray
-    """
-    
-    # Iterate over columns from the end
-    for idx in range(arr.shape[1]-1, -1, -1):
-        if np.any(arr[:, idx]):
-            return arr[:, :(idx+1)]
-    return np.empty((arr.shape[0], 0))
-
-
-import os
-
-def create_directory_for_filepath(filepath: str) -> None:
-    """
-    Given a file path, creates the underlying directory structure if it doesn't already exist.
-
-    Args:
-        filepath (str): The path to the file for which the directory structure should be created.
-
-    Raises:
-        ValueError: If the provided path is empty or None.
-        OSError: If there's an error creating the directory structure.
-    """
-    
-    if not filepath:
-        raise ValueError("The provided filepath is empty or None.")
-    
-    directory = os.path.dirname(filepath)
-    
-    if directory and not os.path.exists(directory):
-        try:
-            os.makedirs(directory)
-            print(f"Directory structure {directory} created successfully.")
-        except OSError as e:
-            raise OSError(f"Error creating directory structure {directory}. Error: {e}")
-
-# Example usage:
-# create_directory_for_filepath("/path/to/directory/that/might/not/exist/filename.txt")
-
-def check_file_exists(file_path: str) -> bool:
-    """
-    Checks if the provided file path exists.
-
-    Args:
-        file_path (str): Path to the file.
-
-    Returns:
-        bool: True if the file exists, raises ValueError otherwise.
-    """
-    if os.path.exists(file_path):
-        return True
-    else:
-        raise ValueError(f"The provided file path '{file_path}' does not exist.")
-
-def count_gpus():
-    # Count NVIDIA GPUs
-    import torch
-    nvidia_gpu_count = torch.cuda.device_count()
-
-    # Count AMD GPUs
-    amd_gpu_count = 0
-    try:
-        clinfo_output = subprocess.check_output('clinfo').decode('utf-8')
-        amd_gpu_count = clinfo_output.count('Device Type: GPU')
-    except:
-        pass  # clinfo command might not be available
-
-    total_gpus = nvidia_gpu_count + amd_gpu_count
-
-    return total_gpus
-
-
-def create_hard_links(source_directory: str, target_directory: str, blacklist: list = []) -> None:
-    """
-    Creates hard links for all files from the source directory to the target directory.
-    
-    Args:
-        source_directory (str): The directory containing the original files.
-        target_directory (str): The directory where hard links will be created.
-        blacklist (list): List of filenames to exclude from creating hard links.
-    
-    Returns:
-        None
-    """
-    
-    # Ensure the provided directories exist
-    if not os.path.exists(source_directory):
-        raise ValueError(f"The source directory '{source_directory}' does not exist.")
-    if not os.path.exists(target_directory):
-        os.makedirs(target_directory)
-    
-    # Iterate through the files in the source directory
-    for filename in os.listdir(source_directory):
-        source_file_path = os.path.join(source_directory, filename)
-        target_file_path = os.path.join(target_directory, filename)
-        
-        # Check for files to skip
-        if (filename.startswith('.') or 
-            filename.startswith('_') or 
-            os.path.isdir(source_file_path) or 
-            filename in blacklist):
-            continue
-        
-        # Create a hard link
-        os.link(source_file_path, target_file_path)
-
-    return f"Hard links created in {target_directory} from {source_directory}."
-
-# Example usage
-# create_hard_links("/path/to/source_directory", "/path/to/target_directory", blacklist=["file_to_skip.txt"])
-
-def create_selected_hard_links(source_directory: str, target_directory: str, filenames: list) -> None:
-    """
-    Creates hard links for the specified files from the source directory to the target directory.
-    
-    Args:
-        source_directory (str): The directory containing the original files.
-        target_directory (str): The directory where hard links will be created.
-        filenames (list): List of filenames for which hard links should be created.
-    
-    Returns:
-        None
-    """
-    
-    # Ensure the provided directories exist
-    if not os.path.exists(source_directory):
-        raise ValueError(f"The source directory '{source_directory}' does not exist.")
-    if not os.path.exists(target_directory):
-        os.makedirs(target_directory)
-    
-    # Iterate through the specified filenames
-    for filename in filenames:
-        source_file_path = os.path.join(source_directory, filename)
-        target_file_path = os.path.join(target_directory, filename)
-        
-        # Ensure the file exists in the source directory
-        if not os.path.isfile(source_file_path):
-            print(f"Warning: {filename} does not exist in the source directory. Skipping.")
-            continue
-        
-        # Create a hard link
-        try:
-            os.link(source_file_path, target_file_path)
-        except FileExistsError:
-            print(f'The target hard link {target_file_path} exist. Skipping...')
-
-    return f"Hard links for specified files created in {target_directory} from {source_directory}."
-
-def remove_hidden_files(directory: str) -> None:
-    """
-    Removes all files recursively in a folder that start with '.' or '_'.
-    
-    Args:
-        directory (str): The directory from which hidden files should be removed.
-    
-    Returns:
-        None
-    """
-    
-    # Ensure the directory exists
-    if not os.path.exists(directory):
-        raise ValueError(f"The directory '{directory}' does not exist.")
-    
-    # Use os.walk to iterate through all subdirectories and files
-    for dirpath, dirnames, filenames in os.walk(directory, topdown=False):
-        
-        # Filter out directories starting with '.' or '_'
-        dirnames[:] = [d for d in dirnames if not d.startswith('.') and not d.startswith('_')]
-        
-        # Remove files starting with '.' or '_'
-        for filename in filenames:
-            if filename.startswith('.') or filename.startswith('_'):
-                file_path = os.path.join(dirpath, filename)
-                os.remove(file_path)
-                print(f"Removed: {file_path}")
-                
-    print(f"All hidden files removed from {directory}.")

generation_config.json

@@ -1,5 +1,5 @@
 {
   "_from_model_config": true,
   "pad_token_id": 0,
-  "transformers_version": "4.30.2"
+  "transformers_version": "4.33.2"
 }

prokbert_tokenizer.py

@@ -1,429 +0,0 @@
-# coding=utf-8
-# Based on https://github.com/huggingface/transformers/blob/main/src/transformers/models/bert/tokenization_bert.py
-
-
-# ProkBERT tokenizer stuff
-
-import collections
-import os
-import unicodedata
-from typing import List, Optional, Tuple, Union
-from copy import deepcopy
-from transformers import PreTrainedTokenizer
-from transformers.tokenization_utils import _is_control, _is_punctuation, _is_whitespace
-from transformers.utils import logging
-
-# These utils contains the tools needed by the ProkBERT tokenizer
-
-from config_utils import *
-from sequtils import *
-
-import logging as logger
-
-#logger = logging.get_logger(__name__)
-
-VOCAB_FILES_NAMES = {"vocab_file": "vocab.txt"}
-
-# models prokbert-mini-k6s1, prokbert-large-k6s2, prokbert-large-k6s1
-
-
-PRETRAINED_VOCAB_FILES_MAP = {
-    "vocab_file": {
-        "prokbert-mini-k6s1": "prokbert-base-dna6/vocab.txt",
-        "prokbert-large-k6s1": "prokbert-base-dna6/vocab.txt",
-        "prokbert-large-k6s2": "prokbert-base-dna6/vocab.txt"
-    }
-}
-
-
-PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {
-    "prokbert-mini-k6s1": 1024,
-    "prokbert-large-k6s1": 1024,
-    "prokbert-large-k6s2": 1024
-}
-
-PRETRAINED_INIT_CONFIGURATION = {
-    "prokbert-mini-k6s1": {"do_upper_case": True},
-    "prokbert-large-k6s1": {"do_upper_case": True},
-    "prokbert-large-k6s2": {"do_upper_case": True}
-
-}
-
-
-def load_vocab(vocab_file):
-    """Loads a vocabulary file into a dictionary."""
-    vocab = collections.OrderedDict()
-    with open(vocab_file, "r", encoding="utf-8") as reader:
-        tokens = reader.readlines()
-    for index, token in enumerate(tokens):
-        token = token.rstrip("\n")
-        vocab[token] = index
-    return vocab
-
-
-class ProkBERTTokenizer(PreTrainedTokenizer):
-    """Custom tokenizer for ProkBERT."""
-    
-    vocab_files_names = VOCAB_FILES_NAMES
-    pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
-    pretrained_init_configuration = PRETRAINED_INIT_CONFIGURATION
-    max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
-    nucleotide_abc = {'A', 'T', 'C', 'G'}
-    extended_nucleotide_abc = {'A', 'T', 'C', 'G', '*'}
-    sequence_unk_token = 'N'
-    default_unk_token="[UNK]"
-    default_sep_token="[SEP]"
-    default_pad_token="[PAD]"
-    default_cls_token="[CLS]"
-    default_mask_token="[MASK]"
-
-
-    def __init__(self, 
-                 tokenization_params: Dict = {},
-                 segmentation_params: Dict = {},
-                 comp_params: Dict = {},
-                 operation_space: str = 'sequence',
-                 **kwargs):
-        """Initialize the ProkBERT tokenizer.
-        
-        Args:
-            tokenization_params (Dict, optional): Tokenization parameters. Defaults to {}.
-            segmentation_params (Dict, optional): Segmentation parameters. Defaults to {}.
-            comp_params (Dict, optional): Computational parameters. Defaults to {}.
-            operation_space (str, optional): Specifies the operation mode. Can be 'kmer' or 'sequence'. Defaults to 'kmer'.
-        """
-        super().__init__(cls_token=ProkBERTTokenizer.default_cls_token,
-                         **kwargs)
-        
-        self.defconfig = SeqConfig()
-        self.tokenization_params = self.defconfig.get_and_set_tokenization_parameters(tokenization_params)
-        self.segmentation_params = self.defconfig.get_and_set_segmentation_parameters(segmentation_params)
-        self.comp_params = self.defconfig.get_and_set_computational_parameters(comp_params)
-        self.operation_space = operation_space
-
-        vocab_file = self.tokenization_params['vocabfile']
-        self.vocab = self.tokenization_params['vocabmap']
-        self.id2token = {v: k for k, v in self.vocab.items()}
-        self.max_len = self.tokenization_params['max_segment_length']
-
-        if self.operation_space == 'sequence':
-            token_extension = sorted(list(set(generate_kmers(ProkBERTTokenizer.extended_nucleotide_abc, self.tokenization_params['kmer'])) - \
-                 set(generate_kmers(ProkBERTTokenizer.nucleotide_abc, self.tokenization_params['kmer'])) ))
-            self.extended_vocab = deepcopy(self.vocab)
-            for token in token_extension:
-                self.extended_vocab[token] = 4
-                            
-            self.unk_token = ProkBERTTokenizer.sequence_unk_token * self.tokenization_params['shift']
-            self.mask_token = '*'
-            self.extended_vocab[self.mask_token] = self.vocab['[MASK]']
-
-            full_unk = 'N' * self.tokenization_params['kmer']
-            self.vocab[full_unk] = 1
-            self.id2token[1] = full_unk
-            self.full_unk_token = full_unk
-
-        else:
-            self.extended_vocab = self.vocab 
-            self.unk_token = '[UNK]'
-        self.sep_token = '[SEP]'
-        self.cls_token = '[CLS]'
-        self.pad_token = '[PAD]'
-        self.mask_token = '[MASK]'
-        self.special_tokens = list(self.special_tokens_map.values())
-
-    def __len__(self) -> int:
-        return len(self.vocab)
-
-
-    def tokenize(self, text: str, lca_shift: int = 0, all: bool = False) -> Union[List[str], Tuple[List[List[str]], List[List[str]]]]:
-        """
-        Tokenizes a given segment.
-
-        Args:
-            text (str): The DNA segment to tokenize.
-            lca_shift (int, optional): Which tokenized vector belonging to the specified LCA offset should be returned. Defaults to 0.
-            all (bool, optional): If True, returns all possible tokenizations. Defaults to False.
-        
-        Returns:
-            Union[List[str], Tuple[List[List[str]], List[List[str]]]]: Tokenized segment or tuple of all possible tokenizations.
-        
-        Usage Example:
-            >>> tokenizer = ProkBERTTokenizer(...)
-            >>> segment = 'AATCAAGGAATTATTATCGTT'
-            >>> tokens, kmers = tokenizer.tokenize(segment, all=True)
-            >>> print(tokens)
-            ...
-        """
-        tokenized_segments, kmerized_segments = lca_tokenize_segment(text, self.tokenization_params)
-        if all:
-            return tokenized_segments, kmerized_segments
-        else:
-            return kmerized_segments[lca_shift]
-
-    def _convert_token_to_id(self, token):
-        """Converts a token (str) in an id using the vocab."""
-        return self.vocab.get(token, self.vocab.get(self.unk_token))
-
-    def _convert_id_to_token(self, index):
-        """Converts an index (integer) in a token (str) using the vocab."""
-        return self.ids_to_tokens.get(index, self.unk_token)
-            
-
-    def depr_convert_ids_to_tokens(self, ids: Union[int, List[int]]) -> List[str]:
-        """
-        Converts tokens to their corresponding IDs.
-
-        Args:
-            tokens (List[str]): List of tokens to convert.
-        
-        Returns:
-            List[int]: List of corresponding token IDs.
-        
-        Usage Example:
-            >>> tokenizer = ProkBERTTokenizer(...)
-            >>> tokens = ['AATCAA', 'TCAAGG']
-            >>> ids = tokenizer.convert_tokens_to_ids(tokens)
-            >>> print(ids)
-            ...
-        """
-
-        if isinstance(ids, int):
-            token_ids = self.vocab.get(ids, self.vocab[self.unk_token])
-
-
-        if self.operation_space == 'sequence':
-            token_ids = [self.vocab.get(token, self.vocab[self.full_unk_token]) for token in tokens]
-        
-        else:
-            token_ids = [self.vocab.get(token, self.vocab[self.unk_token]) for token in tokens]
-        
-        return token_ids
-
-    def convert_ids_to_tokens(self, ids: Union[int, List[int]]) -> Union[str, List[str]]:
-        """
-        Converts token IDs back to their original tokens.
-        
-        Args:
-            ids (List[int]): List of token IDs to convert.
-        
-        Returns:
-            List[str]: List of corresponding tokens.
-        
-        Usage Example:
-            >>> tokenizer = ProkBERTTokenizer(...)
-            >>> ids = [213, 3343]
-            >>> tokens = tokenizer.convert_ids_to_tokens(ids)
-            >>> print(tokens)
-            ...
-        """
-        if isinstance(ids, int):
-            ids = [ids]
-        if len(ids) == 1: 
-            #default_token_list = [self.id2token.get(ids[0], self.unk_token)]
-            return self.id2token.get(ids[0], self.unk_token)
-
-        if self.operation_space == 'kmer':
-            token_list = [self.id2token.get(id, self.unk_token) for id in ids]
-
-        elif self.operation_space == 'sequence':
-            token_list = []
-            # Handling the sentence start
-            if ids[0] == 2:
-                pass
-            else:
-                token_list.append(self.id2token.get(ids[0], self.unk_token))
-            if len(ids) > 1:
-                # if this is a kmer then we add accordingly. 
-                true_start_token = self.id2token.get(ids[1], self.unk_token)
-
-
-                token_list.append(true_start_token)
-            print(token_list)
-            if len(ids) >2:
-                # Adding the other tokens until the end
-                for token_id in ids[2:]:
-                    mapped_token_id = self.id2token.get(token_id, self.unk_token)
-                    if (mapped_token_id in self.special_tokens):
-                        act_token_value = ''
-                    else:
-                        act_token_value = mapped_token_id[-1*self.tokenization_params['shift']:]
-                        token_list.append(act_token_value)
-
-        return token_list
-    
-    
-    def save_vocabulary(self, save_directory: str, filename_prefix: Optional[str] = None) -> Tuple[str]:
-        """Saves the vocabulary to a file."""
-        if filename_prefix is None:
-            filename_prefix = ""
-        vocab_file_path = os.path.join(save_directory, filename_prefix + "vocab.txt")
-        with open(vocab_file_path, "w") as f:
-            for token in self.vocab:
-                f.write(token + "\\n")
-        return (vocab_file_path,)
-
-    @classmethod
-    def from_pretrained(cls, vocab_file: str) -> 'ProkBERTTokenizer':
-        """Loads a pre-trained tokenizer.
-        
-        Args:
-            vocab_file (str): Path to the pre-trained tokenizer vocabulary file.
-        
-        Returns:
-            ProkBERTTokenizer: Loaded tokenizer instance.
-        """
-        return cls(vocab_file)
-
-    def encode_plus(self, text: str, lca_shift: int = 0, **kwargs) -> Dict[str, np.ndarray]:
-        """
-        Tokenizes a sequence and returns it in a format suitable for model input.
-        
-        Args:
-            text (str): The sequence to tokenize.
-            lca_shift (int, optional): LCA offset for tokenization. Defaults to 0.
-        
-        Returns:
-            Dict[str, np.ndarray]: Dictionary containing token IDs and attention masks.
-        
-        Usage Example:
-            >>> tokenizer = ProkBERTTokenizer(...)
-            >>> segment = 'AATCAAGGAATTATTATCGTT'
-            >>> encoded = tokenizer.encode_plus(segment)
-            >>> print(encoded)
-            ...
-        """
-        tokenized_segments, kmerized_segments = lca_tokenize_segment(text, self.tokenization_params)
-        input_ids = tokenized_segments[lca_shift]
-        attention_mask = [1] * len(input_ids)
-
-        # Padding
-        while len(input_ids) < self.max_len:
-            input_ids.append(0)
-            attention_mask.append(0)
-
-        return {
-            "input_ids": np.array(input_ids, dtype=self.comp_params['np_tokentype']),
-            "attention_mask": np.array(attention_mask, dtype=self.comp_params['np_tokentype'])
-        }
-    
-    def batch_encode_plus(self, sequences: List[str], lca_shift: int = 0, all: bool = False, **kwargs) -> Dict[str, List[List[int]]]:
-        """
-        Tokenizes multiple sequences and returns them in a format suitable for model input. It is assumed that sequences 
-        have already been preprocessed (i.e., segmented) and quality controlled. 
-
-        Args:
-        - sequences (List[str]): A list of DNA sequences to be tokenized.
-        - lca_shift (int, default=0): The LCA offset or windows to get the tokenized vector. If the required offset is >= shift, 
-        an error is raised.
-        - all (bool, default=False): Whether all possible tokenization vectors should be returned. If False, only the specified 
-        offset is used. 
-        - **kwargs: Additional arguments (like max_length, padding, etc.)
-
-        Returns:
-        - Dict[str, List[List[int]]]: A dictionary containing token IDs, attention masks, and token type IDs.
-        """
-        shift = self.tokenization_params['shift']
-        if lca_shift >= shift:
-            raise ValueError(f'The required offset {lca_shift} is invalid. The maximum offset should be < {shift}')
-        
-        # Parallel tokenization. First, create unique IDs for all sequences. 
-        sequence_ids = list(range(len(sequences)))
-        to_tokenize_data = (sequences, sequence_ids)
-        
-        # Tokenize each sequence
-        tokenization_results = batch_tokenize_segments_with_ids(
-            to_tokenize_data, 
-            self.tokenization_params, 
-            self.comp_params['cpu_cores_for_tokenization'],
-            self.comp_params['batch_size_tokenization'],
-            self.comp_params['np_tokentype']
-        )
-
-        # Generate input ids, token type ids, and attention masks
-        input_ids = []
-        token_type_ids = []
-        attention_masks = []
-        
-        if all:
-            for tokenized_vectors in tokenization_results.values():
-                for tokenized_vector in tokenized_vectors:
-                    input_ids.append(tokenized_vector)
-                    token_type_ids.append([0] * len(tokenized_vector))
-                    attention_masks.append([1] * len(tokenized_vector))
-        else:
-            for tokenized_vectors in tokenization_results.values():
-                selected_vector = tokenized_vectors[lca_shift]
-                input_ids.append(selected_vector)
-                token_type_ids.append([0] * len(selected_vector))
-                attention_masks.append([1] * len(selected_vector))
-
-        return {
-            "input_ids": input_ids,
-            "token_type_ids": token_type_ids,
-            "attention_mask": attention_masks
-        }
-    
-    def encode(self, segment: str,  lca_shift: int = 0, all: bool = False, add_special_tokens: bool = True, **kwargs) -> List[int]:
-        """
-        Encode a DNA sequence into its corresponding token IDs.
-        
-        Args:
-            text (str): The DNA segment to encode.
-            add_special_tokens (bool, optional): Whether to add special tokens like [CLS] and [SEP]. Defaults to True.
-        
-        Returns:
-            List[int]: Encoded token IDs.
-        
-        Usage Example:
-            >>> tokenizer = ProkBERTTokenizer(...)
-            >>> segment = 'AATCAAGGAATTATTATCGTT'
-            >>> ids = tokenizer.encode(segment)
-            >>> print(ids)
-            ...
-        """
-        shift = self.tokenization_params['shift']
-        if lca_shift >= shift:
-            raise ValueError(f'The required offset {lca_shift} is invalid. The maximum offset should be < {shift}')
-        
-        tokenized_segments, _ = lca_tokenize_segment(segment, self.tokenization_params)
-
-        # if all is set to True, then we return all the possible ids as a list
-        if all:
-            token_ids = tokenized_segments
-            if not add_special_tokens:
-                new_token_ids = []
-                for token_id_set in tokenized_segments:
-                    new_token_ids.append(token_id_set[1:len(token_id_set)-1])
-                token_ids = new_token_ids
-
-        else:
-            token_ids = tokenized_segments[lca_shift]
-            # Convert tokens to their corresponding IDs
-            # Add special tokens if needed
-            if not add_special_tokens:
-                token_ids = token_ids[1:len(token_ids)-1]
-
-        return token_ids
-    
-    def decode(self, ids):
-        tokens = self.convert_ids_to_tokens(ids)
-        return ''.join(tokens)
-    
-    def batch_decode(self, token_ids_list: List[List[int]], **kwargs) -> List[str]:
-        """
-        Decodes multiple token ID sequences back into their original sequences.
-        
-        Args:
-            token_ids_list (List[List[int]]): List of token ID sequences.
-        
-        Returns:
-            List[str]: List of decoded sequences.
-        
-        Usage Example:
-            >>> tokenizer = ProkBERTTokenizer(...)
-            >>> ids = [[2, 213, 3343, 165, 2580, 248, 3905, 978, 3296, 3]]
-            >>> sequences = tokenizer.batch_decode(ids)
-            >>> print(sequences)
-            ...
-        """
-        return [self.decode(token_ids) for token_ids in token_ids_list]

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c251c7687fddf753b1961ef191957f2229c166f49cd847d7b9fc230ad11b672b
-size 82613521
+oid sha256:a2c0eaf12de5f9e61ec02eff8e331f35efea6673beff819c9c598eb08fc6e8b1
+size 82605450

tokenizer_config.json

@@ -1,6 +1,12 @@
 {
   "clean_up_tokenization_spaces": true,
   "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
   "model_max_length": 1000000000000000019884624838656,
-  "tokenizer_class": "ProkBERTTokenizer"
-}
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "tokenizer_class": "LCATokenizer",
+  "unk_token": "[UNK]",
+  "kmer": 6,
+  "shift": 1
+}

vocab.txt

@@ -1 +1,4102 @@
-[PAD]\n[UNK]\n[CLS]\n[SEP]\n[MASK]\nAAAAAA\nAAAAAC\nAAAAAG\nAAAAAT\nAAAACA\nAAAACC\nAAAACG\nAAAACT\nAAAAGA\nAAAAGC\nAAAAGG\nAAAAGT\nAAAATA\nAAAATC\nAAAATG\nAAAATT\nAAACAA\nAAACAC\nAAACAG\nAAACAT\nAAACCA\nAAACCC\nAAACCG\nAAACCT\nAAACGA\nAAACGC\nAAACGG\nAAACGT\nAAACTA\nAAACTC\nAAACTG\nAAACTT\nAAAGAA\nAAAGAC\nAAAGAG\nAAAGAT\nAAAGCA\nAAAGCC\nAAAGCG\nAAAGCT\nAAAGGA\nAAAGGC\nAAAGGG\nAAAGGT\nAAAGTA\nAAAGTC\nAAAGTG\nAAAGTT\nAAATAA\nAAATAC\nAAATAG\nAAATAT\nAAATCA\nAAATCC\nAAATCG\nAAATCT\nAAATGA\nAAATGC\nAAATGG\nAAATGT\nAAATTA\nAAATTC\nAAATTG\nAAATTT\nAACAAA\nAACAAC\nAACAAG\nAACAAT\nAACACA\nAACACC\nAACACG\nAACACT\nAACAGA\nAACAGC\nAACAGG\nAACAGT\nAACATA\nAACATC\nAACATG\nAACATT\nAACCAA\nAACCAC\nAACCAG\nAACCAT\nAACCCA\nAACCCC\nAACCCG\nAACCCT\nAACCGA\nAACCGC\nAACCGG\nAACCGT\nAACCTA\nAACCTC\nAACCTG\nAACCTT\nAACGAA\nAACGAC\nAACGAG\nAACGAT\nAACGCA\nAACGCC\nAACGCG\nAACGCT\nAACGGA\nAACGGC\nAACGGG\nAACGGT\nAACGTA\nAACGTC\nAACGTG\nAACGTT\nAACTAA\nAACTAC\nAACTAG\nAACTAT\nAACTCA\nAACTCC\nAACTCG\nAACTCT\nAACTGA\nAACTGC\nAACTGG\nAACTGT\nAACTTA\nAACTTC\nAACTTG\nAACTTT\nAAGAAA\nAAGAAC\nAAGAAG\nAAGAAT\nAAGACA\nAAGACC\nAAGACG\nAAGACT\nAAGAGA\nAAGAGC\nAAGAGG\nAAGAGT\nAAGATA\nAAGATC\nAAGATG\nAAGATT\nAAGCAA\nAAGCAC\nAAGCAG\nAAGCAT\nAAGCCA\nAAGCCC\nAAGCCG\nAAGCCT\nAAGCGA\nAAGCGC\nAAGCGG\nAAGCGT\nAAGCTA\nAAGCTC\nAAGCTG\nAAGCTT\nAAGGAA\nAAGGAC\nAAGGAG\nAAGGAT\nAAGGCA\nAAGGCC\nAAGGCG\nAAGGCT\nAAGGGA\nAAGGGC\nAAGGGG\nAAGGGT\nAAGGTA\nAAGGTC\nAAGGTG\nAAGGTT\nAAGTAA\nAAGTAC\nAAGTAG\nAAGTAT\nAAGTCA\nAAGTCC\nAAGTCG\nAAGTCT\nAAGTGA\nAAGTGC\nAAGTGG\nAAGTGT\nAAGTTA\nAAGTTC\nAAGTTG\nAAGTTT\nAATAAA\nAATAAC\nAATAAG\nAATAAT\nAATACA\nAATACC\nAATACG\nAATACT\nAATAGA\nAATAGC\nAATAGG\nAATAGT\nAATATA\nAATATC\nAATATG\nAATATT\nAATCAA\nAATCAC\nAATCAG\nAATCAT\nAATCCA\nAATCCC\nAATCCG\nAATCCT\nAATCGA\nAATCGC\nAATCGG\nAATCGT\nAATCTA\nAATCTC\nAATCTG\nAATCTT\nAATGAA\nAATGAC\nAATGAG\nAATGAT\nAATGCA\nAATGCC\nAATGCG\nAATGCT\nAATGGA\nAATGGC\nAATGGG\nAATGGT\nAATGTA\nAATGTC\nAATGTG\nAATGTT\nAATTAA\nAATTAC\nAATTAG\nAATTAT\nAATTCA\nAATTCC\nAATTCG\nAATTCT\nAATTGA\nAATTGC\nAATTGG\nAATTGT\nAATTTA\nAATTTC\nAATTTG\nAATTTT\nACAAAA\nACAAAC\nACAAAG\nACAAAT\nACAACA\nACAACC\nACAACG\nACAACT\nACAAGA\nACAAGC\nACAAGG\nACAAGT\nACAATA\nACAATC\nACAATG\nACAATT\nACACAA\nACACAC\nACACAG\nACACAT\nACACCA\nACACCC\nACACCG\nACACCT\nACACGA\nACACGC\nACACGG\nACACGT\nACACTA\nACACTC\nACACTG\nACACTT\nACAGAA\nACAGAC\nACAGAG\nACAGAT\nACAGCA\nACAGCC\nACAGCG\nACAGCT\nACAGGA\nACAGGC\nACAGGG\nACAGGT\nACAGTA\nACAGTC\nACAGTG\nACAGTT\nACATAA\nACATAC\nACATAG\nACATAT\nACATCA\nACATCC\nACATCG\nACATCT\nACATGA\nACATGC\nACATGG\nACATGT\nACATTA\nACATTC\nACATTG\nACATTT\nACCAAA\nACCAAC\nACCAAG\nACCAAT\nACCACA\nACCACC\nACCACG\nACCACT\nACCAGA\nACCAGC\nACCAGG\nACCAGT\nACCATA\nACCATC\nACCATG\nACCATT\nACCCAA\nACCCAC\nACCCAG\nACCCAT\nACCCCA\nACCCCC\nACCCCG\nACCCCT\nACCCGA\nACCCGC\nACCCGG\nACCCGT\nACCCTA\nACCCTC\nACCCTG\nACCCTT\nACCGAA\nACCGAC\nACCGAG\nACCGAT\nACCGCA\nACCGCC\nACCGCG\nACCGCT\nACCGGA\nACCGGC\nACCGGG\nACCGGT\nACCGTA\nACCGTC\nACCGTG\nACCGTT\nACCTAA\nACCTAC\nACCTAG\nACCTAT\nACCTCA\nACCTCC\nACCTCG\nACCTCT\nACCTGA\nACCTGC\nACCTGG\nACCTGT\nACCTTA\nACCTTC\nACCTTG\nACCTTT\nACGAAA\nACGAAC\nACGAAG\nACGAAT\nACGACA\nACGACC\nACGACG\nACGACT\nACGAGA\nACGAGC\nACGAGG\nACGAGT\nACGATA\nACGATC\nACGATG\nACGATT\nACGCAA\nACGCAC\nACGCAG\nACGCAT\nACGCCA\nACGCCC\nACGCCG\nACGCCT\nACGCGA\nACGCGC\nACGCGG\nACGCGT\nACGCTA\nACGCTC\nACGCTG\nACGCTT\nACGGAA\nACGGAC\nACGGAG\nACGGAT\nACGGCA\nACGGCC\nACGGCG\nACGGCT\nACGGGA\nACGGGC\nACGGGG\nACGGGT\nACGGTA\nACGGTC\nACGGTG\nACGGTT\nACGTAA\nACGTAC\nACGTAG\nACGTAT\nACGTCA\nACGTCC\nACGTCG\nACGTCT\nACGTGA\nACGTGC\nACGTGG\nACGTGT\nACGTTA\nACGTTC\nACGTTG\nACGTTT\nACTAAA\nACTAAC\nACTAAG\nACTAAT\nACTACA\nACTACC\nACTACG\nACTACT\nACTAGA\nACTAGC\nACTAGG\nACTAGT\nACTATA\nACTATC\nACTATG\nACTATT\nACTCAA\nACTCAC\nACTCAG\nACTCAT\nACTCCA\nACTCCC\nACTCCG\nACTCCT\nACTCGA\nACTCGC\nACTCGG\nACTCGT\nACTCTA\nACTCTC\nACTCTG\nACTCTT\nACTGAA\nACTGAC\nACTGAG\nACTGAT\nACTGCA\nACTGCC\nACTGCG\nACTGCT\nACTGGA\nACTGGC\nACTGGG\nACTGGT\nACTGTA\nACTGTC\nACTGTG\nACTGTT\nACTTAA\nACTTAC\nACTTAG\nACTTAT\nACTTCA\nACTTCC\nACTTCG\nACTTCT\nACTTGA\nACTTGC\nACTTGG\nACTTGT\nACTTTA\nACTTTC\nACTTTG\nACTTTT\nAGAAAA\nAGAAAC\nAGAAAG\nAGAAAT\nAGAACA\nAGAACC\nAGAACG\nAGAACT\nAGAAGA\nAGAAGC\nAGAAGG\nAGAAGT\nAGAATA\nAGAATC\nAGAATG\nAGAATT\nAGACAA\nAGACAC\nAGACAG\nAGACAT\nAGACCA\nAGACCC\nAGACCG\nAGACCT\nAGACGA\nAGACGC\nAGACGG\nAGACGT\nAGACTA\nAGACTC\nAGACTG\nAGACTT\nAGAGAA\nAGAGAC\nAGAGAG\nAGAGAT\nAGAGCA\nAGAGCC\nAGAGCG\nAGAGCT\nAGAGGA\nAGAGGC\nAGAGGG\nAGAGGT\nAGAGTA\nAGAGTC\nAGAGTG\nAGAGTT\nAGATAA\nAGATAC\nAGATAG\nAGATAT\nAGATCA\nAGATCC\nAGATCG\nAGATCT\nAGATGA\nAGATGC\nAGATGG\nAGATGT\nAGATTA\nAGATTC\nAGATTG\nAGATTT\nAGCAAA\nAGCAAC\nAGCAAG\nAGCAAT\nAGCACA\nAGCACC\nAGCACG\nAGCACT\nAGCAGA\nAGCAGC\nAGCAGG\nAGCAGT\nAGCATA\nAGCATC\nAGCATG\nAGCATT\nAGCCAA\nAGCCAC\nAGCCAG\nAGCCAT\nAGCCCA\nAGCCCC\nAGCCCG\nAGCCCT\nAGCCGA\nAGCCGC\nAGCCGG\nAGCCGT\nAGCCTA\nAGCCTC\nAGCCTG\nAGCCTT\nAGCGAA\nAGCGAC\nAGCGAG\nAGCGAT\nAGCGCA\nAGCGCC\nAGCGCG\nAGCGCT\nAGCGGA\nAGCGGC\nAGCGGG\nAGCGGT\nAGCGTA\nAGCGTC\nAGCGTG\nAGCGTT\nAGCTAA\nAGCTAC\nAGCTAG\nAGCTAT\nAGCTCA\nAGCTCC\nAGCTCG\nAGCTCT\nAGCTGA\nAGCTGC\nAGCTGG\nAGCTGT\nAGCTTA\nAGCTTC\nAGCTTG\nAGCTTT\nAGGAAA\nAGGAAC\nAGGAAG\nAGGAAT\nAGGACA\nAGGACC\nAGGACG\nAGGACT\nAGGAGA\nAGGAGC\nAGGAGG\nAGGAGT\nAGGATA\nAGGATC\nAGGATG\nAGGATT\nAGGCAA\nAGGCAC\nAGGCAG\nAGGCAT\nAGGCCA\nAGGCCC\nAGGCCG\nAGGCCT\nAGGCGA\nAGGCGC\nAGGCGG\nAGGCGT\nAGGCTA\nAGGCTC\nAGGCTG\nAGGCTT\nAGGGAA\nAGGGAC\nAGGGAG\nAGGGAT\nAGGGCA\nAGGGCC\nAGGGCG\nAGGGCT\nAGGGGA\nAGGGGC\nAGGGGG\nAGGGGT\nAGGGTA\nAGGGTC\nAGGGTG\nAGGGTT\nAGGTAA\nAGGTAC\nAGGTAG\nAGGTAT\nAGGTCA\nAGGTCC\nAGGTCG\nAGGTCT\nAGGTGA\nAGGTGC\nAGGTGG\nAGGTGT\nAGGTTA\nAGGTTC\nAGGTTG\nAGGTTT\nAGTAAA\nAGTAAC\nAGTAAG\nAGTAAT\nAGTACA\nAGTACC\nAGTACG\nAGTACT\nAGTAGA\nAGTAGC\nAGTAGG\nAGTAGT\nAGTATA\nAGTATC\nAGTATG\nAGTATT\nAGTCAA\nAGTCAC\nAGTCAG\nAGTCAT\nAGTCCA\nAGTCCC\nAGTCCG\nAGTCCT\nAGTCGA\nAGTCGC\nAGTCGG\nAGTCGT\nAGTCTA\nAGTCTC\nAGTCTG\nAGTCTT\nAGTGAA\nAGTGAC\nAGTGAG\nAGTGAT\nAGTGCA\nAGTGCC\nAGTGCG\nAGTGCT\nAGTGGA\nAGTGGC\nAGTGGG\nAGTGGT\nAGTGTA\nAGTGTC\nAGTGTG\nAGTGTT\nAGTTAA\nAGTTAC\nAGTTAG\nAGTTAT\nAGTTCA\nAGTTCC\nAGTTCG\nAGTTCT\nAGTTGA\nAGTTGC\nAGTTGG\nAGTTGT\nAGTTTA\nAGTTTC\nAGTTTG\nAGTTTT\nATAAAA\nATAAAC\nATAAAG\nATAAAT\nATAACA\nATAACC\nATAACG\nATAACT\nATAAGA\nATAAGC\nATAAGG\nATAAGT\nATAATA\nATAATC\nATAATG\nATAATT\nATACAA\nATACAC\nATACAG\nATACAT\nATACCA\nATACCC\nATACCG\nATACCT\nATACGA\nATACGC\nATACGG\nATACGT\nATACTA\nATACTC\nATACTG\nATACTT\nATAGAA\nATAGAC\nATAGAG\nATAGAT\nATAGCA\nATAGCC\nATAGCG\nATAGCT\nATAGGA\nATAGGC\nATAGGG\nATAGGT\nATAGTA\nATAGTC\nATAGTG\nATAGTT\nATATAA\nATATAC\nATATAG\nATATAT\nATATCA\nATATCC\nATATCG\nATATCT\nATATGA\nATATGC\nATATGG\nATATGT\nATATTA\nATATTC\nATATTG\nATATTT\nATCAAA\nATCAAC\nATCAAG\nATCAAT\nATCACA\nATCACC\nATCACG\nATCACT\nATCAGA\nATCAGC\nATCAGG\nATCAGT\nATCATA\nATCATC\nATCATG\nATCATT\nATCCAA\nATCCAC\nATCCAG\nATCCAT\nATCCCA\nATCCCC\nATCCCG\nATCCCT\nATCCGA\nATCCGC\nATCCGG\nATCCGT\nATCCTA\nATCCTC\nATCCTG\nATCCTT\nATCGAA\nATCGAC\nATCGAG\nATCGAT\nATCGCA\nATCGCC\nATCGCG\nATCGCT\nATCGGA\nATCGGC\nATCGGG\nATCGGT\nATCGTA\nATCGTC\nATCGTG\nATCGTT\nATCTAA\nATCTAC\nATCTAG\nATCTAT\nATCTCA\nATCTCC\nATCTCG\nATCTCT\nATCTGA\nATCTGC\nATCTGG\nATCTGT\nATCTTA\nATCTTC\nATCTTG\nATCTTT\nATGAAA\nATGAAC\nATGAAG\nATGAAT\nATGACA\nATGACC\nATGACG\nATGACT\nATGAGA\nATGAGC\nATGAGG\nATGAGT\nATGATA\nATGATC\nATGATG\nATGATT\nATGCAA\nATGCAC\nATGCAG\nATGCAT\nATGCCA\nATGCCC\nATGCCG\nATGCCT\nATGCGA\nATGCGC\nATGCGG\nATGCGT\nATGCTA\nATGCTC\nATGCTG\nATGCTT\nATGGAA\nATGGAC\nATGGAG\nATGGAT\nATGGCA\nATGGCC\nATGGCG\nATGGCT\nATGGGA\nATGGGC\nATGGGG\nATGGGT\nATGGTA\nATGGTC\nATGGTG\nATGGTT\nATGTAA\nATGTAC\nATGTAG\nATGTAT\nATGTCA\nATGTCC\nATGTCG\nATGTCT\nATGTGA\nATGTGC\nATGTGG\nATGTGT\nATGTTA\nATGTTC\nATGTTG\nATGTTT\nATTAAA\nATTAAC\nATTAAG\nATTAAT\nATTACA\nATTACC\nATTACG\nATTACT\nATTAGA\nATTAGC\nATTAGG\nATTAGT\nATTATA\nATTATC\nATTATG\nATTATT\nATTCAA\nATTCAC\nATTCAG\nATTCAT\nATTCCA\nATTCCC\nATTCCG\nATTCCT\nATTCGA\nATTCGC\nATTCGG\nATTCGT\nATTCTA\nATTCTC\nATTCTG\nATTCTT\nATTGAA\nATTGAC\nATTGAG\nATTGAT\nATTGCA\nATTGCC\nATTGCG\nATTGCT\nATTGGA\nATTGGC\nATTGGG\nATTGGT\nATTGTA\nATTGTC\nATTGTG\nATTGTT\nATTTAA\nATTTAC\nATTTAG\nATTTAT\nATTTCA\nATTTCC\nATTTCG\nATTTCT\nATTTGA\nATTTGC\nATTTGG\nATTTGT\nATTTTA\nATTTTC\nATTTTG\nATTTTT\nCAAAAA\nCAAAAC\nCAAAAG\nCAAAAT\nCAAACA\nCAAACC\nCAAACG\nCAAACT\nCAAAGA\nCAAAGC\nCAAAGG\nCAAAGT\nCAAATA\nCAAATC\nCAAATG\nCAAATT\nCAACAA\nCAACAC\nCAACAG\nCAACAT\nCAACCA\nCAACCC\nCAACCG\nCAACCT\nCAACGA\nCAACGC\nCAACGG\nCAACGT\nCAACTA\nCAACTC\nCAACTG\nCAACTT\nCAAGAA\nCAAGAC\nCAAGAG\nCAAGAT\nCAAGCA\nCAAGCC\nCAAGCG\nCAAGCT\nCAAGGA\nCAAGGC\nCAAGGG\nCAAGGT\nCAAGTA\nCAAGTC\nCAAGTG\nCAAGTT\nCAATAA\nCAATAC\nCAATAG\nCAATAT\nCAATCA\nCAATCC\nCAATCG\nCAATCT\nCAATGA\nCAATGC\nCAATGG\nCAATGT\nCAATTA\nCAATTC\nCAATTG\nCAATTT\nCACAAA\nCACAAC\nCACAAG\nCACAAT\nCACACA\nCACACC\nCACACG\nCACACT\nCACAGA\nCACAGC\nCACAGG\nCACAGT\nCACATA\nCACATC\nCACATG\nCACATT\nCACCAA\nCACCAC\nCACCAG\nCACCAT\nCACCCA\nCACCCC\nCACCCG\nCACCCT\nCACCGA\nCACCGC\nCACCGG\nCACCGT\nCACCTA\nCACCTC\nCACCTG\nCACCTT\nCACGAA\nCACGAC\nCACGAG\nCACGAT\nCACGCA\nCACGCC\nCACGCG\nCACGCT\nCACGGA\nCACGGC\nCACGGG\nCACGGT\nCACGTA\nCACGTC\nCACGTG\nCACGTT\nCACTAA\nCACTAC\nCACTAG\nCACTAT\nCACTCA\nCACTCC\nCACTCG\nCACTCT\nCACTGA\nCACTGC\nCACTGG\nCACTGT\nCACTTA\nCACTTC\nCACTTG\nCACTTT\nCAGAAA\nCAGAAC\nCAGAAG\nCAGAAT\nCAGACA\nCAGACC\nCAGACG\nCAGACT\nCAGAGA\nCAGAGC\nCAGAGG\nCAGAGT\nCAGATA\nCAGATC\nCAGATG\nCAGATT\nCAGCAA\nCAGCAC\nCAGCAG\nCAGCAT\nCAGCCA\nCAGCCC\nCAGCCG\nCAGCCT\nCAGCGA\nCAGCGC\nCAGCGG\nCAGCGT\nCAGCTA\nCAGCTC\nCAGCTG\nCAGCTT\nCAGGAA\nCAGGAC\nCAGGAG\nCAGGAT\nCAGGCA\nCAGGCC\nCAGGCG\nCAGGCT\nCAGGGA\nCAGGGC\nCAGGGG\nCAGGGT\nCAGGTA\nCAGGTC\nCAGGTG\nCAGGTT\nCAGTAA\nCAGTAC\nCAGTAG\nCAGTAT\nCAGTCA\nCAGTCC\nCAGTCG\nCAGTCT\nCAGTGA\nCAGTGC\nCAGTGG\nCAGTGT\nCAGTTA\nCAGTTC\nCAGTTG\nCAGTTT\nCATAAA\nCATAAC\nCATAAG\nCATAAT\nCATACA\nCATACC\nCATACG\nCATACT\nCATAGA\nCATAGC\nCATAGG\nCATAGT\nCATATA\nCATATC\nCATATG\nCATATT\nCATCAA\nCATCAC\nCATCAG\nCATCAT\nCATCCA\nCATCCC\nCATCCG\nCATCCT\nCATCGA\nCATCGC\nCATCGG\nCATCGT\nCATCTA\nCATCTC\nCATCTG\nCATCTT\nCATGAA\nCATGAC\nCATGAG\nCATGAT\nCATGCA\nCATGCC\nCATGCG\nCATGCT\nCATGGA\nCATGGC\nCATGGG\nCATGGT\nCATGTA\nCATGTC\nCATGTG\nCATGTT\nCATTAA\nCATTAC\nCATTAG\nCATTAT\nCATTCA\nCATTCC\nCATTCG\nCATTCT\nCATTGA\nCATTGC\nCATTGG\nCATTGT\nCATTTA\nCATTTC\nCATTTG\nCATTTT\nCCAAAA\nCCAAAC\nCCAAAG\nCCAAAT\nCCAACA\nCCAACC\nCCAACG\nCCAACT\nCCAAGA\nCCAAGC\nCCAAGG\nCCAAGT\nCCAATA\nCCAATC\nCCAATG\nCCAATT\nCCACAA\nCCACAC\nCCACAG\nCCACAT\nCCACCA\nCCACCC\nCCACCG\nCCACCT\nCCACGA\nCCACGC\nCCACGG\nCCACGT\nCCACTA\nCCACTC\nCCACTG\nCCACTT\nCCAGAA\nCCAGAC\nCCAGAG\nCCAGAT\nCCAGCA\nCCAGCC\nCCAGCG\nCCAGCT\nCCAGGA\nCCAGGC\nCCAGGG\nCCAGGT\nCCAGTA\nCCAGTC\nCCAGTG\nCCAGTT\nCCATAA\nCCATAC\nCCATAG\nCCATAT\nCCATCA\nCCATCC\nCCATCG\nCCATCT\nCCATGA\nCCATGC\nCCATGG\nCCATGT\nCCATTA\nCCATTC\nCCATTG\nCCATTT\nCCCAAA\nCCCAAC\nCCCAAG\nCCCAAT\nCCCACA\nCCCACC\nCCCACG\nCCCACT\nCCCAGA\nCCCAGC\nCCCAGG\nCCCAGT\nCCCATA\nCCCATC\nCCCATG\nCCCATT\nCCCCAA\nCCCCAC\nCCCCAG\nCCCCAT\nCCCCCA\nCCCCCC\nCCCCCG\nCCCCCT\nCCCCGA\nCCCCGC\nCCCCGG\nCCCCGT\nCCCCTA\nCCCCTC\nCCCCTG\nCCCCTT\nCCCGAA\nCCCGAC\nCCCGAG\nCCCGAT\nCCCGCA\nCCCGCC\nCCCGCG\nCCCGCT\nCCCGGA\nCCCGGC\nCCCGGG\nCCCGGT\nCCCGTA\nCCCGTC\nCCCGTG\nCCCGTT\nCCCTAA\nCCCTAC\nCCCTAG\nCCCTAT\nCCCTCA\nCCCTCC\nCCCTCG\nCCCTCT\nCCCTGA\nCCCTGC\nCCCTGG\nCCCTGT\nCCCTTA\nCCCTTC\nCCCTTG\nCCCTTT\nCCGAAA\nCCGAAC\nCCGAAG\nCCGAAT\nCCGACA\nCCGACC\nCCGACG\nCCGACT\nCCGAGA\nCCGAGC\nCCGAGG\nCCGAGT\nCCGATA\nCCGATC\nCCGATG\nCCGATT\nCCGCAA\nCCGCAC\nCCGCAG\nCCGCAT\nCCGCCA\nCCGCCC\nCCGCCG\nCCGCCT\nCCGCGA\nCCGCGC\nCCGCGG\nCCGCGT\nCCGCTA\nCCGCTC\nCCGCTG\nCCGCTT\nCCGGAA\nCCGGAC\nCCGGAG\nCCGGAT\nCCGGCA\nCCGGCC\nCCGGCG\nCCGGCT\nCCGGGA\nCCGGGC\nCCGGGG\nCCGGGT\nCCGGTA\nCCGGTC\nCCGGTG\nCCGGTT\nCCGTAA\nCCGTAC\nCCGTAG\nCCGTAT\nCCGTCA\nCCGTCC\nCCGTCG\nCCGTCT\nCCGTGA\nCCGTGC\nCCGTGG\nCCGTGT\nCCGTTA\nCCGTTC\nCCGTTG\nCCGTTT\nCCTAAA\nCCTAAC\nCCTAAG\nCCTAAT\nCCTACA\nCCTACC\nCCTACG\nCCTACT\nCCTAGA\nCCTAGC\nCCTAGG\nCCTAGT\nCCTATA\nCCTATC\nCCTATG\nCCTATT\nCCTCAA\nCCTCAC\nCCTCAG\nCCTCAT\nCCTCCA\nCCTCCC\nCCTCCG\nCCTCCT\nCCTCGA\nCCTCGC\nCCTCGG\nCCTCGT\nCCTCTA\nCCTCTC\nCCTCTG\nCCTCTT\nCCTGAA\nCCTGAC\nCCTGAG\nCCTGAT\nCCTGCA\nCCTGCC\nCCTGCG\nCCTGCT\nCCTGGA\nCCTGGC\nCCTGGG\nCCTGGT\nCCTGTA\nCCTGTC\nCCTGTG\nCCTGTT\nCCTTAA\nCCTTAC\nCCTTAG\nCCTTAT\nCCTTCA\nCCTTCC\nCCTTCG\nCCTTCT\nCCTTGA\nCCTTGC\nCCTTGG\nCCTTGT\nCCTTTA\nCCTTTC\nCCTTTG\nCCTTTT\nCGAAAA\nCGAAAC\nCGAAAG\nCGAAAT\nCGAACA\nCGAACC\nCGAACG\nCGAACT\nCGAAGA\nCGAAGC\nCGAAGG\nCGAAGT\nCGAATA\nCGAATC\nCGAATG\nCGAATT\nCGACAA\nCGACAC\nCGACAG\nCGACAT\nCGACCA\nCGACCC\nCGACCG\nCGACCT\nCGACGA\nCGACGC\nCGACGG\nCGACGT\nCGACTA\nCGACTC\nCGACTG\nCGACTT\nCGAGAA\nCGAGAC\nCGAGAG\nCGAGAT\nCGAGCA\nCGAGCC\nCGAGCG\nCGAGCT\nCGAGGA\nCGAGGC\nCGAGGG\nCGAGGT\nCGAGTA\nCGAGTC\nCGAGTG\nCGAGTT\nCGATAA\nCGATAC\nCGATAG\nCGATAT\nCGATCA\nCGATCC\nCGATCG\nCGATCT\nCGATGA\nCGATGC\nCGATGG\nCGATGT\nCGATTA\nCGATTC\nCGATTG\nCGATTT\nCGCAAA\nCGCAAC\nCGCAAG\nCGCAAT\nCGCACA\nCGCACC\nCGCACG\nCGCACT\nCGCAGA\nCGCAGC\nCGCAGG\nCGCAGT\nCGCATA\nCGCATC\nCGCATG\nCGCATT\nCGCCAA\nCGCCAC\nCGCCAG\nCGCCAT\nCGCCCA\nCGCCCC\nCGCCCG\nCGCCCT\nCGCCGA\nCGCCGC\nCGCCGG\nCGCCGT\nCGCCTA\nCGCCTC\nCGCCTG\nCGCCTT\nCGCGAA\nCGCGAC\nCGCGAG\nCGCGAT\nCGCGCA\nCGCGCC\nCGCGCG\nCGCGCT\nCGCGGA\nCGCGGC\nCGCGGG\nCGCGGT\nCGCGTA\nCGCGTC\nCGCGTG\nCGCGTT\nCGCTAA\nCGCTAC\nCGCTAG\nCGCTAT\nCGCTCA\nCGCTCC\nCGCTCG\nCGCTCT\nCGCTGA\nCGCTGC\nCGCTGG\nCGCTGT\nCGCTTA\nCGCTTC\nCGCTTG\nCGCTTT\nCGGAAA\nCGGAAC\nCGGAAG\nCGGAAT\nCGGACA\nCGGACC\nCGGACG\nCGGACT\nCGGAGA\nCGGAGC\nCGGAGG\nCGGAGT\nCGGATA\nCGGATC\nCGGATG\nCGGATT\nCGGCAA\nCGGCAC\nCGGCAG\nCGGCAT\nCGGCCA\nCGGCCC\nCGGCCG\nCGGCCT\nCGGCGA\nCGGCGC\nCGGCGG\nCGGCGT\nCGGCTA\nCGGCTC\nCGGCTG\nCGGCTT\nCGGGAA\nCGGGAC\nCGGGAG\nCGGGAT\nCGGGCA\nCGGGCC\nCGGGCG\nCGGGCT\nCGGGGA\nCGGGGC\nCGGGGG\nCGGGGT\nCGGGTA\nCGGGTC\nCGGGTG\nCGGGTT\nCGGTAA\nCGGTAC\nCGGTAG\nCGGTAT\nCGGTCA\nCGGTCC\nCGGTCG\nCGGTCT\nCGGTGA\nCGGTGC\nCGGTGG\nCGGTGT\nCGGTTA\nCGGTTC\nCGGTTG\nCGGTTT\nCGTAAA\nCGTAAC\nCGTAAG\nCGTAAT\nCGTACA\nCGTACC\nCGTACG\nCGTACT\nCGTAGA\nCGTAGC\nCGTAGG\nCGTAGT\nCGTATA\nCGTATC\nCGTATG\nCGTATT\nCGTCAA\nCGTCAC\nCGTCAG\nCGTCAT\nCGTCCA\nCGTCCC\nCGTCCG\nCGTCCT\nCGTCGA\nCGTCGC\nCGTCGG\nCGTCGT\nCGTCTA\nCGTCTC\nCGTCTG\nCGTCTT\nCGTGAA\nCGTGAC\nCGTGAG\nCGTGAT\nCGTGCA\nCGTGCC\nCGTGCG\nCGTGCT\nCGTGGA\nCGTGGC\nCGTGGG\nCGTGGT\nCGTGTA\nCGTGTC\nCGTGTG\nCGTGTT\nCGTTAA\nCGTTAC\nCGTTAG\nCGTTAT\nCGTTCA\nCGTTCC\nCGTTCG\nCGTTCT\nCGTTGA\nCGTTGC\nCGTTGG\nCGTTGT\nCGTTTA\nCGTTTC\nCGTTTG\nCGTTTT\nCTAAAA\nCTAAAC\nCTAAAG\nCTAAAT\nCTAACA\nCTAACC\nCTAACG\nCTAACT\nCTAAGA\nCTAAGC\nCTAAGG\nCTAAGT\nCTAATA\nCTAATC\nCTAATG\nCTAATT\nCTACAA\nCTACAC\nCTACAG\nCTACAT\nCTACCA\nCTACCC\nCTACCG\nCTACCT\nCTACGA\nCTACGC\nCTACGG\nCTACGT\nCTACTA\nCTACTC\nCTACTG\nCTACTT\nCTAGAA\nCTAGAC\nCTAGAG\nCTAGAT\nCTAGCA\nCTAGCC\nCTAGCG\nCTAGCT\nCTAGGA\nCTAGGC\nCTAGGG\nCTAGGT\nCTAGTA\nCTAGTC\nCTAGTG\nCTAGTT\nCTATAA\nCTATAC\nCTATAG\nCTATAT\nCTATCA\nCTATCC\nCTATCG\nCTATCT\nCTATGA\nCTATGC\nCTATGG\nCTATGT\nCTATTA\nCTATTC\nCTATTG\nCTATTT\nCTCAAA\nCTCAAC\nCTCAAG\nCTCAAT\nCTCACA\nCTCACC\nCTCACG\nCTCACT\nCTCAGA\nCTCAGC\nCTCAGG\nCTCAGT\nCTCATA\nCTCATC\nCTCATG\nCTCATT\nCTCCAA\nCTCCAC\nCTCCAG\nCTCCAT\nCTCCCA\nCTCCCC\nCTCCCG\nCTCCCT\nCTCCGA\nCTCCGC\nCTCCGG\nCTCCGT\nCTCCTA\nCTCCTC\nCTCCTG\nCTCCTT\nCTCGAA\nCTCGAC\nCTCGAG\nCTCGAT\nCTCGCA\nCTCGCC\nCTCGCG\nCTCGCT\nCTCGGA\nCTCGGC\nCTCGGG\nCTCGGT\nCTCGTA\nCTCGTC\nCTCGTG\nCTCGTT\nCTCTAA\nCTCTAC\nCTCTAG\nCTCTAT\nCTCTCA\nCTCTCC\nCTCTCG\nCTCTCT\nCTCTGA\nCTCTGC\nCTCTGG\nCTCTGT\nCTCTTA\nCTCTTC\nCTCTTG\nCTCTTT\nCTGAAA\nCTGAAC\nCTGAAG\nCTGAAT\nCTGACA\nCTGACC\nCTGACG\nCTGACT\nCTGAGA\nCTGAGC\nCTGAGG\nCTGAGT\nCTGATA\nCTGATC\nCTGATG\nCTGATT\nCTGCAA\nCTGCAC\nCTGCAG\nCTGCAT\nCTGCCA\nCTGCCC\nCTGCCG\nCTGCCT\nCTGCGA\nCTGCGC\nCTGCGG\nCTGCGT\nCTGCTA\nCTGCTC\nCTGCTG\nCTGCTT\nCTGGAA\nCTGGAC\nCTGGAG\nCTGGAT\nCTGGCA\nCTGGCC\nCTGGCG\nCTGGCT\nCTGGGA\nCTGGGC\nCTGGGG\nCTGGGT\nCTGGTA\nCTGGTC\nCTGGTG\nCTGGTT\nCTGTAA\nCTGTAC\nCTGTAG\nCTGTAT\nCTGTCA\nCTGTCC\nCTGTCG\nCTGTCT\nCTGTGA\nCTGTGC\nCTGTGG\nCTGTGT\nCTGTTA\nCTGTTC\nCTGTTG\nCTGTTT\nCTTAAA\nCTTAAC\nCTTAAG\nCTTAAT\nCTTACA\nCTTACC\nCTTACG\nCTTACT\nCTTAGA\nCTTAGC\nCTTAGG\nCTTAGT\nCTTATA\nCTTATC\nCTTATG\nCTTATT\nCTTCAA\nCTTCAC\nCTTCAG\nCTTCAT\nCTTCCA\nCTTCCC\nCTTCCG\nCTTCCT\nCTTCGA\nCTTCGC\nCTTCGG\nCTTCGT\nCTTCTA\nCTTCTC\nCTTCTG\nCTTCTT\nCTTGAA\nCTTGAC\nCTTGAG\nCTTGAT\nCTTGCA\nCTTGCC\nCTTGCG\nCTTGCT\nCTTGGA\nCTTGGC\nCTTGGG\nCTTGGT\nCTTGTA\nCTTGTC\nCTTGTG\nCTTGTT\nCTTTAA\nCTTTAC\nCTTTAG\nCTTTAT\nCTTTCA\nCTTTCC\nCTTTCG\nCTTTCT\nCTTTGA\nCTTTGC\nCTTTGG\nCTTTGT\nCTTTTA\nCTTTTC\nCTTTTG\nCTTTTT\nGAAAAA\nGAAAAC\nGAAAAG\nGAAAAT\nGAAACA\nGAAACC\nGAAACG\nGAAACT\nGAAAGA\nGAAAGC\nGAAAGG\nGAAAGT\nGAAATA\nGAAATC\nGAAATG\nGAAATT\nGAACAA\nGAACAC\nGAACAG\nGAACAT\nGAACCA\nGAACCC\nGAACCG\nGAACCT\nGAACGA\nGAACGC\nGAACGG\nGAACGT\nGAACTA\nGAACTC\nGAACTG\nGAACTT\nGAAGAA\nGAAGAC\nGAAGAG\nGAAGAT\nGAAGCA\nGAAGCC\nGAAGCG\nGAAGCT\nGAAGGA\nGAAGGC\nGAAGGG\nGAAGGT\nGAAGTA\nGAAGTC\nGAAGTG\nGAAGTT\nGAATAA\nGAATAC\nGAATAG\nGAATAT\nGAATCA\nGAATCC\nGAATCG\nGAATCT\nGAATGA\nGAATGC\nGAATGG\nGAATGT\nGAATTA\nGAATTC\nGAATTG\nGAATTT\nGACAAA\nGACAAC\nGACAAG\nGACAAT\nGACACA\nGACACC\nGACACG\nGACACT\nGACAGA\nGACAGC\nGACAGG\nGACAGT\nGACATA\nGACATC\nGACATG\nGACATT\nGACCAA\nGACCAC\nGACCAG\nGACCAT\nGACCCA\nGACCCC\nGACCCG\nGACCCT\nGACCGA\nGACCGC\nGACCGG\nGACCGT\nGACCTA\nGACCTC\nGACCTG\nGACCTT\nGACGAA\nGACGAC\nGACGAG\nGACGAT\nGACGCA\nGACGCC\nGACGCG\nGACGCT\nGACGGA\nGACGGC\nGACGGG\nGACGGT\nGACGTA\nGACGTC\nGACGTG\nGACGTT\nGACTAA\nGACTAC\nGACTAG\nGACTAT\nGACTCA\nGACTCC\nGACTCG\nGACTCT\nGACTGA\nGACTGC\nGACTGG\nGACTGT\nGACTTA\nGACTTC\nGACTTG\nGACTTT\nGAGAAA\nGAGAAC\nGAGAAG\nGAGAAT\nGAGACA\nGAGACC\nGAGACG\nGAGACT\nGAGAGA\nGAGAGC\nGAGAGG\nGAGAGT\nGAGATA\nGAGATC\nGAGATG\nGAGATT\nGAGCAA\nGAGCAC\nGAGCAG\nGAGCAT\nGAGCCA\nGAGCCC\nGAGCCG\nGAGCCT\nGAGCGA\nGAGCGC\nGAGCGG\nGAGCGT\nGAGCTA\nGAGCTC\nGAGCTG\nGAGCTT\nGAGGAA\nGAGGAC\nGAGGAG\nGAGGAT\nGAGGCA\nGAGGCC\nGAGGCG\nGAGGCT\nGAGGGA\nGAGGGC\nGAGGGG\nGAGGGT\nGAGGTA\nGAGGTC\nGAGGTG\nGAGGTT\nGAGTAA\nGAGTAC\nGAGTAG\nGAGTAT\nGAGTCA\nGAGTCC\nGAGTCG\nGAGTCT\nGAGTGA\nGAGTGC\nGAGTGG\nGAGTGT\nGAGTTA\nGAGTTC\nGAGTTG\nGAGTTT\nGATAAA\nGATAAC\nGATAAG\nGATAAT\nGATACA\nGATACC\nGATACG\nGATACT\nGATAGA\nGATAGC\nGATAGG\nGATAGT\nGATATA\nGATATC\nGATATG\nGATATT\nGATCAA\nGATCAC\nGATCAG\nGATCAT\nGATCCA\nGATCCC\nGATCCG\nGATCCT\nGATCGA\nGATCGC\nGATCGG\nGATCGT\nGATCTA\nGATCTC\nGATCTG\nGATCTT\nGATGAA\nGATGAC\nGATGAG\nGATGAT\nGATGCA\nGATGCC\nGATGCG\nGATGCT\nGATGGA\nGATGGC\nGATGGG\nGATGGT\nGATGTA\nGATGTC\nGATGTG\nGATGTT\nGATTAA\nGATTAC\nGATTAG\nGATTAT\nGATTCA\nGATTCC\nGATTCG\nGATTCT\nGATTGA\nGATTGC\nGATTGG\nGATTGT\nGATTTA\nGATTTC\nGATTTG\nGATTTT\nGCAAAA\nGCAAAC\nGCAAAG\nGCAAAT\nGCAACA\nGCAACC\nGCAACG\nGCAACT\nGCAAGA\nGCAAGC\nGCAAGG\nGCAAGT\nGCAATA\nGCAATC\nGCAATG\nGCAATT\nGCACAA\nGCACAC\nGCACAG\nGCACAT\nGCACCA\nGCACCC\nGCACCG\nGCACCT\nGCACGA\nGCACGC\nGCACGG\nGCACGT\nGCACTA\nGCACTC\nGCACTG\nGCACTT\nGCAGAA\nGCAGAC\nGCAGAG\nGCAGAT\nGCAGCA\nGCAGCC\nGCAGCG\nGCAGCT\nGCAGGA\nGCAGGC\nGCAGGG\nGCAGGT\nGCAGTA\nGCAGTC\nGCAGTG\nGCAGTT\nGCATAA\nGCATAC\nGCATAG\nGCATAT\nGCATCA\nGCATCC\nGCATCG\nGCATCT\nGCATGA\nGCATGC\nGCATGG\nGCATGT\nGCATTA\nGCATTC\nGCATTG\nGCATTT\nGCCAAA\nGCCAAC\nGCCAAG\nGCCAAT\nGCCACA\nGCCACC\nGCCACG\nGCCACT\nGCCAGA\nGCCAGC\nGCCAGG\nGCCAGT\nGCCATA\nGCCATC\nGCCATG\nGCCATT\nGCCCAA\nGCCCAC\nGCCCAG\nGCCCAT\nGCCCCA\nGCCCCC\nGCCCCG\nGCCCCT\nGCCCGA\nGCCCGC\nGCCCGG\nGCCCGT\nGCCCTA\nGCCCTC\nGCCCTG\nGCCCTT\nGCCGAA\nGCCGAC\nGCCGAG\nGCCGAT\nGCCGCA\nGCCGCC\nGCCGCG\nGCCGCT\nGCCGGA\nGCCGGC\nGCCGGG\nGCCGGT\nGCCGTA\nGCCGTC\nGCCGTG\nGCCGTT\nGCCTAA\nGCCTAC\nGCCTAG\nGCCTAT\nGCCTCA\nGCCTCC\nGCCTCG\nGCCTCT\nGCCTGA\nGCCTGC\nGCCTGG\nGCCTGT\nGCCTTA\nGCCTTC\nGCCTTG\nGCCTTT\nGCGAAA\nGCGAAC\nGCGAAG\nGCGAAT\nGCGACA\nGCGACC\nGCGACG\nGCGACT\nGCGAGA\nGCGAGC\nGCGAGG\nGCGAGT\nGCGATA\nGCGATC\nGCGATG\nGCGATT\nGCGCAA\nGCGCAC\nGCGCAG\nGCGCAT\nGCGCCA\nGCGCCC\nGCGCCG\nGCGCCT\nGCGCGA\nGCGCGC\nGCGCGG\nGCGCGT\nGCGCTA\nGCGCTC\nGCGCTG\nGCGCTT\nGCGGAA\nGCGGAC\nGCGGAG\nGCGGAT\nGCGGCA\nGCGGCC\nGCGGCG\nGCGGCT\nGCGGGA\nGCGGGC\nGCGGGG\nGCGGGT\nGCGGTA\nGCGGTC\nGCGGTG\nGCGGTT\nGCGTAA\nGCGTAC\nGCGTAG\nGCGTAT\nGCGTCA\nGCGTCC\nGCGTCG\nGCGTCT\nGCGTGA\nGCGTGC\nGCGTGG\nGCGTGT\nGCGTTA\nGCGTTC\nGCGTTG\nGCGTTT\nGCTAAA\nGCTAAC\nGCTAAG\nGCTAAT\nGCTACA\nGCTACC\nGCTACG\nGCTACT\nGCTAGA\nGCTAGC\nGCTAGG\nGCTAGT\nGCTATA\nGCTATC\nGCTATG\nGCTATT\nGCTCAA\nGCTCAC\nGCTCAG\nGCTCAT\nGCTCCA\nGCTCCC\nGCTCCG\nGCTCCT\nGCTCGA\nGCTCGC\nGCTCGG\nGCTCGT\nGCTCTA\nGCTCTC\nGCTCTG\nGCTCTT\nGCTGAA\nGCTGAC\nGCTGAG\nGCTGAT\nGCTGCA\nGCTGCC\nGCTGCG\nGCTGCT\nGCTGGA\nGCTGGC\nGCTGGG\nGCTGGT\nGCTGTA\nGCTGTC\nGCTGTG\nGCTGTT\nGCTTAA\nGCTTAC\nGCTTAG\nGCTTAT\nGCTTCA\nGCTTCC\nGCTTCG\nGCTTCT\nGCTTGA\nGCTTGC\nGCTTGG\nGCTTGT\nGCTTTA\nGCTTTC\nGCTTTG\nGCTTTT\nGGAAAA\nGGAAAC\nGGAAAG\nGGAAAT\nGGAACA\nGGAACC\nGGAACG\nGGAACT\nGGAAGA\nGGAAGC\nGGAAGG\nGGAAGT\nGGAATA\nGGAATC\nGGAATG\nGGAATT\nGGACAA\nGGACAC\nGGACAG\nGGACAT\nGGACCA\nGGACCC\nGGACCG\nGGACCT\nGGACGA\nGGACGC\nGGACGG\nGGACGT\nGGACTA\nGGACTC\nGGACTG\nGGACTT\nGGAGAA\nGGAGAC\nGGAGAG\nGGAGAT\nGGAGCA\nGGAGCC\nGGAGCG\nGGAGCT\nGGAGGA\nGGAGGC\nGGAGGG\nGGAGGT\nGGAGTA\nGGAGTC\nGGAGTG\nGGAGTT\nGGATAA\nGGATAC\nGGATAG\nGGATAT\nGGATCA\nGGATCC\nGGATCG\nGGATCT\nGGATGA\nGGATGC\nGGATGG\nGGATGT\nGGATTA\nGGATTC\nGGATTG\nGGATTT\nGGCAAA\nGGCAAC\nGGCAAG\nGGCAAT\nGGCACA\nGGCACC\nGGCACG\nGGCACT\nGGCAGA\nGGCAGC\nGGCAGG\nGGCAGT\nGGCATA\nGGCATC\nGGCATG\nGGCATT\nGGCCAA\nGGCCAC\nGGCCAG\nGGCCAT\nGGCCCA\nGGCCCC\nGGCCCG\nGGCCCT\nGGCCGA\nGGCCGC\nGGCCGG\nGGCCGT\nGGCCTA\nGGCCTC\nGGCCTG\nGGCCTT\nGGCGAA\nGGCGAC\nGGCGAG\nGGCGAT\nGGCGCA\nGGCGCC\nGGCGCG\nGGCGCT\nGGCGGA\nGGCGGC\nGGCGGG\nGGCGGT\nGGCGTA\nGGCGTC\nGGCGTG\nGGCGTT\nGGCTAA\nGGCTAC\nGGCTAG\nGGCTAT\nGGCTCA\nGGCTCC\nGGCTCG\nGGCTCT\nGGCTGA\nGGCTGC\nGGCTGG\nGGCTGT\nGGCTTA\nGGCTTC\nGGCTTG\nGGCTTT\nGGGAAA\nGGGAAC\nGGGAAG\nGGGAAT\nGGGACA\nGGGACC\nGGGACG\nGGGACT\nGGGAGA\nGGGAGC\nGGGAGG\nGGGAGT\nGGGATA\nGGGATC\nGGGATG\nGGGATT\nGGGCAA\nGGGCAC\nGGGCAG\nGGGCAT\nGGGCCA\nGGGCCC\nGGGCCG\nGGGCCT\nGGGCGA\nGGGCGC\nGGGCGG\nGGGCGT\nGGGCTA\nGGGCTC\nGGGCTG\nGGGCTT\nGGGGAA\nGGGGAC\nGGGGAG\nGGGGAT\nGGGGCA\nGGGGCC\nGGGGCG\nGGGGCT\nGGGGGA\nGGGGGC\nGGGGGG\nGGGGGT\nGGGGTA\nGGGGTC\nGGGGTG\nGGGGTT\nGGGTAA\nGGGTAC\nGGGTAG\nGGGTAT\nGGGTCA\nGGGTCC\nGGGTCG\nGGGTCT\nGGGTGA\nGGGTGC\nGGGTGG\nGGGTGT\nGGGTTA\nGGGTTC\nGGGTTG\nGGGTTT\nGGTAAA\nGGTAAC\nGGTAAG\nGGTAAT\nGGTACA\nGGTACC\nGGTACG\nGGTACT\nGGTAGA\nGGTAGC\nGGTAGG\nGGTAGT\nGGTATA\nGGTATC\nGGTATG\nGGTATT\nGGTCAA\nGGTCAC\nGGTCAG\nGGTCAT\nGGTCCA\nGGTCCC\nGGTCCG\nGGTCCT\nGGTCGA\nGGTCGC\nGGTCGG\nGGTCGT\nGGTCTA\nGGTCTC\nGGTCTG\nGGTCTT\nGGTGAA\nGGTGAC\nGGTGAG\nGGTGAT\nGGTGCA\nGGTGCC\nGGTGCG\nGGTGCT\nGGTGGA\nGGTGGC\nGGTGGG\nGGTGGT\nGGTGTA\nGGTGTC\nGGTGTG\nGGTGTT\nGGTTAA\nGGTTAC\nGGTTAG\nGGTTAT\nGGTTCA\nGGTTCC\nGGTTCG\nGGTTCT\nGGTTGA\nGGTTGC\nGGTTGG\nGGTTGT\nGGTTTA\nGGTTTC\nGGTTTG\nGGTTTT\nGTAAAA\nGTAAAC\nGTAAAG\nGTAAAT\nGTAACA\nGTAACC\nGTAACG\nGTAACT\nGTAAGA\nGTAAGC\nGTAAGG\nGTAAGT\nGTAATA\nGTAATC\nGTAATG\nGTAATT\nGTACAA\nGTACAC\nGTACAG\nGTACAT\nGTACCA\nGTACCC\nGTACCG\nGTACCT\nGTACGA\nGTACGC\nGTACGG\nGTACGT\nGTACTA\nGTACTC\nGTACTG\nGTACTT\nGTAGAA\nGTAGAC\nGTAGAG\nGTAGAT\nGTAGCA\nGTAGCC\nGTAGCG\nGTAGCT\nGTAGGA\nGTAGGC\nGTAGGG\nGTAGGT\nGTAGTA\nGTAGTC\nGTAGTG\nGTAGTT\nGTATAA\nGTATAC\nGTATAG\nGTATAT\nGTATCA\nGTATCC\nGTATCG\nGTATCT\nGTATGA\nGTATGC\nGTATGG\nGTATGT\nGTATTA\nGTATTC\nGTATTG\nGTATTT\nGTCAAA\nGTCAAC\nGTCAAG\nGTCAAT\nGTCACA\nGTCACC\nGTCACG\nGTCACT\nGTCAGA\nGTCAGC\nGTCAGG\nGTCAGT\nGTCATA\nGTCATC\nGTCATG\nGTCATT\nGTCCAA\nGTCCAC\nGTCCAG\nGTCCAT\nGTCCCA\nGTCCCC\nGTCCCG\nGTCCCT\nGTCCGA\nGTCCGC\nGTCCGG\nGTCCGT\nGTCCTA\nGTCCTC\nGTCCTG\nGTCCTT\nGTCGAA\nGTCGAC\nGTCGAG\nGTCGAT\nGTCGCA\nGTCGCC\nGTCGCG\nGTCGCT\nGTCGGA\nGTCGGC\nGTCGGG\nGTCGGT\nGTCGTA\nGTCGTC\nGTCGTG\nGTCGTT\nGTCTAA\nGTCTAC\nGTCTAG\nGTCTAT\nGTCTCA\nGTCTCC\nGTCTCG\nGTCTCT\nGTCTGA\nGTCTGC\nGTCTGG\nGTCTGT\nGTCTTA\nGTCTTC\nGTCTTG\nGTCTTT\nGTGAAA\nGTGAAC\nGTGAAG\nGTGAAT\nGTGACA\nGTGACC\nGTGACG\nGTGACT\nGTGAGA\nGTGAGC\nGTGAGG\nGTGAGT\nGTGATA\nGTGATC\nGTGATG\nGTGATT\nGTGCAA\nGTGCAC\nGTGCAG\nGTGCAT\nGTGCCA\nGTGCCC\nGTGCCG\nGTGCCT\nGTGCGA\nGTGCGC\nGTGCGG\nGTGCGT\nGTGCTA\nGTGCTC\nGTGCTG\nGTGCTT\nGTGGAA\nGTGGAC\nGTGGAG\nGTGGAT\nGTGGCA\nGTGGCC\nGTGGCG\nGTGGCT\nGTGGGA\nGTGGGC\nGTGGGG\nGTGGGT\nGTGGTA\nGTGGTC\nGTGGTG\nGTGGTT\nGTGTAA\nGTGTAC\nGTGTAG\nGTGTAT\nGTGTCA\nGTGTCC\nGTGTCG\nGTGTCT\nGTGTGA\nGTGTGC\nGTGTGG\nGTGTGT\nGTGTTA\nGTGTTC\nGTGTTG\nGTGTTT\nGTTAAA\nGTTAAC\nGTTAAG\nGTTAAT\nGTTACA\nGTTACC\nGTTACG\nGTTACT\nGTTAGA\nGTTAGC\nGTTAGG\nGTTAGT\nGTTATA\nGTTATC\nGTTATG\nGTTATT\nGTTCAA\nGTTCAC\nGTTCAG\nGTTCAT\nGTTCCA\nGTTCCC\nGTTCCG\nGTTCCT\nGTTCGA\nGTTCGC\nGTTCGG\nGTTCGT\nGTTCTA\nGTTCTC\nGTTCTG\nGTTCTT\nGTTGAA\nGTTGAC\nGTTGAG\nGTTGAT\nGTTGCA\nGTTGCC\nGTTGCG\nGTTGCT\nGTTGGA\nGTTGGC\nGTTGGG\nGTTGGT\nGTTGTA\nGTTGTC\nGTTGTG\nGTTGTT\nGTTTAA\nGTTTAC\nGTTTAG\nGTTTAT\nGTTTCA\nGTTTCC\nGTTTCG\nGTTTCT\nGTTTGA\nGTTTGC\nGTTTGG\nGTTTGT\nGTTTTA\nGTTTTC\nGTTTTG\nGTTTTT\nTAAAAA\nTAAAAC\nTAAAAG\nTAAAAT\nTAAACA\nTAAACC\nTAAACG\nTAAACT\nTAAAGA\nTAAAGC\nTAAAGG\nTAAAGT\nTAAATA\nTAAATC\nTAAATG\nTAAATT\nTAACAA\nTAACAC\nTAACAG\nTAACAT\nTAACCA\nTAACCC\nTAACCG\nTAACCT\nTAACGA\nTAACGC\nTAACGG\nTAACGT\nTAACTA\nTAACTC\nTAACTG\nTAACTT\nTAAGAA\nTAAGAC\nTAAGAG\nTAAGAT\nTAAGCA\nTAAGCC\nTAAGCG\nTAAGCT\nTAAGGA\nTAAGGC\nTAAGGG\nTAAGGT\nTAAGTA\nTAAGTC\nTAAGTG\nTAAGTT\nTAATAA\nTAATAC\nTAATAG\nTAATAT\nTAATCA\nTAATCC\nTAATCG\nTAATCT\nTAATGA\nTAATGC\nTAATGG\nTAATGT\nTAATTA\nTAATTC\nTAATTG\nTAATTT\nTACAAA\nTACAAC\nTACAAG\nTACAAT\nTACACA\nTACACC\nTACACG\nTACACT\nTACAGA\nTACAGC\nTACAGG\nTACAGT\nTACATA\nTACATC\nTACATG\nTACATT\nTACCAA\nTACCAC\nTACCAG\nTACCAT\nTACCCA\nTACCCC\nTACCCG\nTACCCT\nTACCGA\nTACCGC\nTACCGG\nTACCGT\nTACCTA\nTACCTC\nTACCTG\nTACCTT\nTACGAA\nTACGAC\nTACGAG\nTACGAT\nTACGCA\nTACGCC\nTACGCG\nTACGCT\nTACGGA\nTACGGC\nTACGGG\nTACGGT\nTACGTA\nTACGTC\nTACGTG\nTACGTT\nTACTAA\nTACTAC\nTACTAG\nTACTAT\nTACTCA\nTACTCC\nTACTCG\nTACTCT\nTACTGA\nTACTGC\nTACTGG\nTACTGT\nTACTTA\nTACTTC\nTACTTG\nTACTTT\nTAGAAA\nTAGAAC\nTAGAAG\nTAGAAT\nTAGACA\nTAGACC\nTAGACG\nTAGACT\nTAGAGA\nTAGAGC\nTAGAGG\nTAGAGT\nTAGATA\nTAGATC\nTAGATG\nTAGATT\nTAGCAA\nTAGCAC\nTAGCAG\nTAGCAT\nTAGCCA\nTAGCCC\nTAGCCG\nTAGCCT\nTAGCGA\nTAGCGC\nTAGCGG\nTAGCGT\nTAGCTA\nTAGCTC\nTAGCTG\nTAGCTT\nTAGGAA\nTAGGAC\nTAGGAG\nTAGGAT\nTAGGCA\nTAGGCC\nTAGGCG\nTAGGCT\nTAGGGA\nTAGGGC\nTAGGGG\nTAGGGT\nTAGGTA\nTAGGTC\nTAGGTG\nTAGGTT\nTAGTAA\nTAGTAC\nTAGTAG\nTAGTAT\nTAGTCA\nTAGTCC\nTAGTCG\nTAGTCT\nTAGTGA\nTAGTGC\nTAGTGG\nTAGTGT\nTAGTTA\nTAGTTC\nTAGTTG\nTAGTTT\nTATAAA\nTATAAC\nTATAAG\nTATAAT\nTATACA\nTATACC\nTATACG\nTATACT\nTATAGA\nTATAGC\nTATAGG\nTATAGT\nTATATA\nTATATC\nTATATG\nTATATT\nTATCAA\nTATCAC\nTATCAG\nTATCAT\nTATCCA\nTATCCC\nTATCCG\nTATCCT\nTATCGA\nTATCGC\nTATCGG\nTATCGT\nTATCTA\nTATCTC\nTATCTG\nTATCTT\nTATGAA\nTATGAC\nTATGAG\nTATGAT\nTATGCA\nTATGCC\nTATGCG\nTATGCT\nTATGGA\nTATGGC\nTATGGG\nTATGGT\nTATGTA\nTATGTC\nTATGTG\nTATGTT\nTATTAA\nTATTAC\nTATTAG\nTATTAT\nTATTCA\nTATTCC\nTATTCG\nTATTCT\nTATTGA\nTATTGC\nTATTGG\nTATTGT\nTATTTA\nTATTTC\nTATTTG\nTATTTT\nTCAAAA\nTCAAAC\nTCAAAG\nTCAAAT\nTCAACA\nTCAACC\nTCAACG\nTCAACT\nTCAAGA\nTCAAGC\nTCAAGG\nTCAAGT\nTCAATA\nTCAATC\nTCAATG\nTCAATT\nTCACAA\nTCACAC\nTCACAG\nTCACAT\nTCACCA\nTCACCC\nTCACCG\nTCACCT\nTCACGA\nTCACGC\nTCACGG\nTCACGT\nTCACTA\nTCACTC\nTCACTG\nTCACTT\nTCAGAA\nTCAGAC\nTCAGAG\nTCAGAT\nTCAGCA\nTCAGCC\nTCAGCG\nTCAGCT\nTCAGGA\nTCAGGC\nTCAGGG\nTCAGGT\nTCAGTA\nTCAGTC\nTCAGTG\nTCAGTT\nTCATAA\nTCATAC\nTCATAG\nTCATAT\nTCATCA\nTCATCC\nTCATCG\nTCATCT\nTCATGA\nTCATGC\nTCATGG\nTCATGT\nTCATTA\nTCATTC\nTCATTG\nTCATTT\nTCCAAA\nTCCAAC\nTCCAAG\nTCCAAT\nTCCACA\nTCCACC\nTCCACG\nTCCACT\nTCCAGA\nTCCAGC\nTCCAGG\nTCCAGT\nTCCATA\nTCCATC\nTCCATG\nTCCATT\nTCCCAA\nTCCCAC\nTCCCAG\nTCCCAT\nTCCCCA\nTCCCCC\nTCCCCG\nTCCCCT\nTCCCGA\nTCCCGC\nTCCCGG\nTCCCGT\nTCCCTA\nTCCCTC\nTCCCTG\nTCCCTT\nTCCGAA\nTCCGAC\nTCCGAG\nTCCGAT\nTCCGCA\nTCCGCC\nTCCGCG\nTCCGCT\nTCCGGA\nTCCGGC\nTCCGGG\nTCCGGT\nTCCGTA\nTCCGTC\nTCCGTG\nTCCGTT\nTCCTAA\nTCCTAC\nTCCTAG\nTCCTAT\nTCCTCA\nTCCTCC\nTCCTCG\nTCCTCT\nTCCTGA\nTCCTGC\nTCCTGG\nTCCTGT\nTCCTTA\nTCCTTC\nTCCTTG\nTCCTTT\nTCGAAA\nTCGAAC\nTCGAAG\nTCGAAT\nTCGACA\nTCGACC\nTCGACG\nTCGACT\nTCGAGA\nTCGAGC\nTCGAGG\nTCGAGT\nTCGATA\nTCGATC\nTCGATG\nTCGATT\nTCGCAA\nTCGCAC\nTCGCAG\nTCGCAT\nTCGCCA\nTCGCCC\nTCGCCG\nTCGCCT\nTCGCGA\nTCGCGC\nTCGCGG\nTCGCGT\nTCGCTA\nTCGCTC\nTCGCTG\nTCGCTT\nTCGGAA\nTCGGAC\nTCGGAG\nTCGGAT\nTCGGCA\nTCGGCC\nTCGGCG\nTCGGCT\nTCGGGA\nTCGGGC\nTCGGGG\nTCGGGT\nTCGGTA\nTCGGTC\nTCGGTG\nTCGGTT\nTCGTAA\nTCGTAC\nTCGTAG\nTCGTAT\nTCGTCA\nTCGTCC\nTCGTCG\nTCGTCT\nTCGTGA\nTCGTGC\nTCGTGG\nTCGTGT\nTCGTTA\nTCGTTC\nTCGTTG\nTCGTTT\nTCTAAA\nTCTAAC\nTCTAAG\nTCTAAT\nTCTACA\nTCTACC\nTCTACG\nTCTACT\nTCTAGA\nTCTAGC\nTCTAGG\nTCTAGT\nTCTATA\nTCTATC\nTCTATG\nTCTATT\nTCTCAA\nTCTCAC\nTCTCAG\nTCTCAT\nTCTCCA\nTCTCCC\nTCTCCG\nTCTCCT\nTCTCGA\nTCTCGC\nTCTCGG\nTCTCGT\nTCTCTA\nTCTCTC\nTCTCTG\nTCTCTT\nTCTGAA\nTCTGAC\nTCTGAG\nTCTGAT\nTCTGCA\nTCTGCC\nTCTGCG\nTCTGCT\nTCTGGA\nTCTGGC\nTCTGGG\nTCTGGT\nTCTGTA\nTCTGTC\nTCTGTG\nTCTGTT\nTCTTAA\nTCTTAC\nTCTTAG\nTCTTAT\nTCTTCA\nTCTTCC\nTCTTCG\nTCTTCT\nTCTTGA\nTCTTGC\nTCTTGG\nTCTTGT\nTCTTTA\nTCTTTC\nTCTTTG\nTCTTTT\nTGAAAA\nTGAAAC\nTGAAAG\nTGAAAT\nTGAACA\nTGAACC\nTGAACG\nTGAACT\nTGAAGA\nTGAAGC\nTGAAGG\nTGAAGT\nTGAATA\nTGAATC\nTGAATG\nTGAATT\nTGACAA\nTGACAC\nTGACAG\nTGACAT\nTGACCA\nTGACCC\nTGACCG\nTGACCT\nTGACGA\nTGACGC\nTGACGG\nTGACGT\nTGACTA\nTGACTC\nTGACTG\nTGACTT\nTGAGAA\nTGAGAC\nTGAGAG\nTGAGAT\nTGAGCA\nTGAGCC\nTGAGCG\nTGAGCT\nTGAGGA\nTGAGGC\nTGAGGG\nTGAGGT\nTGAGTA\nTGAGTC\nTGAGTG\nTGAGTT\nTGATAA\nTGATAC\nTGATAG\nTGATAT\nTGATCA\nTGATCC\nTGATCG\nTGATCT\nTGATGA\nTGATGC\nTGATGG\nTGATGT\nTGATTA\nTGATTC\nTGATTG\nTGATTT\nTGCAAA\nTGCAAC\nTGCAAG\nTGCAAT\nTGCACA\nTGCACC\nTGCACG\nTGCACT\nTGCAGA\nTGCAGC\nTGCAGG\nTGCAGT\nTGCATA\nTGCATC\nTGCATG\nTGCATT\nTGCCAA\nTGCCAC\nTGCCAG\nTGCCAT\nTGCCCA\nTGCCCC\nTGCCCG\nTGCCCT\nTGCCGA\nTGCCGC\nTGCCGG\nTGCCGT\nTGCCTA\nTGCCTC\nTGCCTG\nTGCCTT\nTGCGAA\nTGCGAC\nTGCGAG\nTGCGAT\nTGCGCA\nTGCGCC\nTGCGCG\nTGCGCT\nTGCGGA\nTGCGGC\nTGCGGG\nTGCGGT\nTGCGTA\nTGCGTC\nTGCGTG\nTGCGTT\nTGCTAA\nTGCTAC\nTGCTAG\nTGCTAT\nTGCTCA\nTGCTCC\nTGCTCG\nTGCTCT\nTGCTGA\nTGCTGC\nTGCTGG\nTGCTGT\nTGCTTA\nTGCTTC\nTGCTTG\nTGCTTT\nTGGAAA\nTGGAAC\nTGGAAG\nTGGAAT\nTGGACA\nTGGACC\nTGGACG\nTGGACT\nTGGAGA\nTGGAGC\nTGGAGG\nTGGAGT\nTGGATA\nTGGATC\nTGGATG\nTGGATT\nTGGCAA\nTGGCAC\nTGGCAG\nTGGCAT\nTGGCCA\nTGGCCC\nTGGCCG\nTGGCCT\nTGGCGA\nTGGCGC\nTGGCGG\nTGGCGT\nTGGCTA\nTGGCTC\nTGGCTG\nTGGCTT\nTGGGAA\nTGGGAC\nTGGGAG\nTGGGAT\nTGGGCA\nTGGGCC\nTGGGCG\nTGGGCT\nTGGGGA\nTGGGGC\nTGGGGG\nTGGGGT\nTGGGTA\nTGGGTC\nTGGGTG\nTGGGTT\nTGGTAA\nTGGTAC\nTGGTAG\nTGGTAT\nTGGTCA\nTGGTCC\nTGGTCG\nTGGTCT\nTGGTGA\nTGGTGC\nTGGTGG\nTGGTGT\nTGGTTA\nTGGTTC\nTGGTTG\nTGGTTT\nTGTAAA\nTGTAAC\nTGTAAG\nTGTAAT\nTGTACA\nTGTACC\nTGTACG\nTGTACT\nTGTAGA\nTGTAGC\nTGTAGG\nTGTAGT\nTGTATA\nTGTATC\nTGTATG\nTGTATT\nTGTCAA\nTGTCAC\nTGTCAG\nTGTCAT\nTGTCCA\nTGTCCC\nTGTCCG\nTGTCCT\nTGTCGA\nTGTCGC\nTGTCGG\nTGTCGT\nTGTCTA\nTGTCTC\nTGTCTG\nTGTCTT\nTGTGAA\nTGTGAC\nTGTGAG\nTGTGAT\nTGTGCA\nTGTGCC\nTGTGCG\nTGTGCT\nTGTGGA\nTGTGGC\nTGTGGG\nTGTGGT\nTGTGTA\nTGTGTC\nTGTGTG\nTGTGTT\nTGTTAA\nTGTTAC\nTGTTAG\nTGTTAT\nTGTTCA\nTGTTCC\nTGTTCG\nTGTTCT\nTGTTGA\nTGTTGC\nTGTTGG\nTGTTGT\nTGTTTA\nTGTTTC\nTGTTTG\nTGTTTT\nTTAAAA\nTTAAAC\nTTAAAG\nTTAAAT\nTTAACA\nTTAACC\nTTAACG\nTTAACT\nTTAAGA\nTTAAGC\nTTAAGG\nTTAAGT\nTTAATA\nTTAATC\nTTAATG\nTTAATT\nTTACAA\nTTACAC\nTTACAG\nTTACAT\nTTACCA\nTTACCC\nTTACCG\nTTACCT\nTTACGA\nTTACGC\nTTACGG\nTTACGT\nTTACTA\nTTACTC\nTTACTG\nTTACTT\nTTAGAA\nTTAGAC\nTTAGAG\nTTAGAT\nTTAGCA\nTTAGCC\nTTAGCG\nTTAGCT\nTTAGGA\nTTAGGC\nTTAGGG\nTTAGGT\nTTAGTA\nTTAGTC\nTTAGTG\nTTAGTT\nTTATAA\nTTATAC\nTTATAG\nTTATAT\nTTATCA\nTTATCC\nTTATCG\nTTATCT\nTTATGA\nTTATGC\nTTATGG\nTTATGT\nTTATTA\nTTATTC\nTTATTG\nTTATTT\nTTCAAA\nTTCAAC\nTTCAAG\nTTCAAT\nTTCACA\nTTCACC\nTTCACG\nTTCACT\nTTCAGA\nTTCAGC\nTTCAGG\nTTCAGT\nTTCATA\nTTCATC\nTTCATG\nTTCATT\nTTCCAA\nTTCCAC\nTTCCAG\nTTCCAT\nTTCCCA\nTTCCCC\nTTCCCG\nTTCCCT\nTTCCGA\nTTCCGC\nTTCCGG\nTTCCGT\nTTCCTA\nTTCCTC\nTTCCTG\nTTCCTT\nTTCGAA\nTTCGAC\nTTCGAG\nTTCGAT\nTTCGCA\nTTCGCC\nTTCGCG\nTTCGCT\nTTCGGA\nTTCGGC\nTTCGGG\nTTCGGT\nTTCGTA\nTTCGTC\nTTCGTG\nTTCGTT\nTTCTAA\nTTCTAC\nTTCTAG\nTTCTAT\nTTCTCA\nTTCTCC\nTTCTCG\nTTCTCT\nTTCTGA\nTTCTGC\nTTCTGG\nTTCTGT\nTTCTTA\nTTCTTC\nTTCTTG\nTTCTTT\nTTGAAA\nTTGAAC\nTTGAAG\nTTGAAT\nTTGACA\nTTGACC\nTTGACG\nTTGACT\nTTGAGA\nTTGAGC\nTTGAGG\nTTGAGT\nTTGATA\nTTGATC\nTTGATG\nTTGATT\nTTGCAA\nTTGCAC\nTTGCAG\nTTGCAT\nTTGCCA\nTTGCCC\nTTGCCG\nTTGCCT\nTTGCGA\nTTGCGC\nTTGCGG\nTTGCGT\nTTGCTA\nTTGCTC\nTTGCTG\nTTGCTT\nTTGGAA\nTTGGAC\nTTGGAG\nTTGGAT\nTTGGCA\nTTGGCC\nTTGGCG\nTTGGCT\nTTGGGA\nTTGGGC\nTTGGGG\nTTGGGT\nTTGGTA\nTTGGTC\nTTGGTG\nTTGGTT\nTTGTAA\nTTGTAC\nTTGTAG\nTTGTAT\nTTGTCA\nTTGTCC\nTTGTCG\nTTGTCT\nTTGTGA\nTTGTGC\nTTGTGG\nTTGTGT\nTTGTTA\nTTGTTC\nTTGTTG\nTTGTTT\nTTTAAA\nTTTAAC\nTTTAAG\nTTTAAT\nTTTACA\nTTTACC\nTTTACG\nTTTACT\nTTTAGA\nTTTAGC\nTTTAGG\nTTTAGT\nTTTATA\nTTTATC\nTTTATG\nTTTATT\nTTTCAA\nTTTCAC\nTTTCAG\nTTTCAT\nTTTCCA\nTTTCCC\nTTTCCG\nTTTCCT\nTTTCGA\nTTTCGC\nTTTCGG\nTTTCGT\nTTTCTA\nTTTCTC\nTTTCTG\nTTTCTT\nTTTGAA\nTTTGAC\nTTTGAG\nTTTGAT\nTTTGCA\nTTTGCC\nTTTGCG\nTTTGCT\nTTTGGA\nTTTGGC\nTTTGGG\nTTTGGT\nTTTGTA\nTTTGTC\nTTTGTG\nTTTGTT\nTTTTAA\nTTTTAC\nTTTTAG\nTTTTAT\nTTTTCA\nTTTTCC\nTTTTCG\nTTTTCT\nTTTTGA\nTTTTGC\nTTTTGG\nTTTTGT\nTTTTTA\nTTTTTC\nTTTTTG\nTTTTTT\nNNNNNN\n
+[PAD]
+[UNK]
+[CLS]
+[SEP]
+[MASK]
+AAAAAA
+AAAAAC
+AAAAAG
+AAAAAT
+AAAACA
+AAAACC
+AAAACG
+AAAACT
+AAAAGA
+AAAAGC
+AAAAGG
+AAAAGT
+AAAATA
+AAAATC
+AAAATG
+AAAATT
+AAACAA
+AAACAC
+AAACAG
+AAACAT
+AAACCA
+AAACCC
+AAACCG
+AAACCT
+AAACGA
+AAACGC
+AAACGG
+AAACGT
+AAACTA
+AAACTC
+AAACTG
+AAACTT
+AAAGAA
+AAAGAC
+AAAGAG
+AAAGAT
+AAAGCA
+AAAGCC
+AAAGCG
+AAAGCT
+AAAGGA
+AAAGGC
+AAAGGG
+AAAGGT
+AAAGTA
+AAAGTC
+AAAGTG
+AAAGTT
+AAATAA
+AAATAC
+AAATAG
+AAATAT
+AAATCA
+AAATCC
+AAATCG
+AAATCT
+AAATGA
+AAATGC
+AAATGG
+AAATGT
+AAATTA
+AAATTC
+AAATTG
+AAATTT
+AACAAA
+AACAAC
+AACAAG
+AACAAT
+AACACA
+AACACC
+AACACG
+AACACT
+AACAGA
+AACAGC
+AACAGG
+AACAGT
+AACATA
+AACATC
+AACATG
+AACATT
+AACCAA
+AACCAC
+AACCAG
+AACCAT
+AACCCA
+AACCCC
+AACCCG
+AACCCT
+AACCGA
+AACCGC
+AACCGG
+AACCGT
+AACCTA
+AACCTC
+AACCTG
+AACCTT
+AACGAA
+AACGAC
+AACGAG
+AACGAT
+AACGCA
+AACGCC
+AACGCG
+AACGCT
+AACGGA
+AACGGC
+AACGGG
+AACGGT
+AACGTA
+AACGTC
+AACGTG
+AACGTT
+AACTAA
+AACTAC
+AACTAG
+AACTAT
+AACTCA
+AACTCC
+AACTCG
+AACTCT
+AACTGA
+AACTGC
+AACTGG
+AACTGT
+AACTTA
+AACTTC
+AACTTG
+AACTTT
+AAGAAA
+AAGAAC
+AAGAAG
+AAGAAT
+AAGACA
+AAGACC
+AAGACG
+AAGACT
+AAGAGA
+AAGAGC
+AAGAGG
+AAGAGT
+AAGATA
+AAGATC
+AAGATG
+AAGATT
+AAGCAA
+AAGCAC
+AAGCAG
+AAGCAT
+AAGCCA
+AAGCCC
+AAGCCG
+AAGCCT
+AAGCGA
+AAGCGC
+AAGCGG
+AAGCGT
+AAGCTA
+AAGCTC
+AAGCTG
+AAGCTT
+AAGGAA
+AAGGAC
+AAGGAG
+AAGGAT
+AAGGCA
+AAGGCC
+AAGGCG
+AAGGCT
+AAGGGA
+AAGGGC
+AAGGGG
+AAGGGT
+AAGGTA
+AAGGTC
+AAGGTG
+AAGGTT
+AAGTAA
+AAGTAC
+AAGTAG
+AAGTAT
+AAGTCA
+AAGTCC
+AAGTCG
+AAGTCT
+AAGTGA
+AAGTGC
+AAGTGG
+AAGTGT
+AAGTTA
+AAGTTC
+AAGTTG
+AAGTTT
+AATAAA
+AATAAC
+AATAAG
+AATAAT
+AATACA
+AATACC
+AATACG
+AATACT
+AATAGA
+AATAGC
+AATAGG
+AATAGT
+AATATA
+AATATC
+AATATG
+AATATT
+AATCAA
+AATCAC
+AATCAG
+AATCAT
+AATCCA
+AATCCC
+AATCCG
+AATCCT
+AATCGA
+AATCGC
+AATCGG
+AATCGT
+AATCTA
+AATCTC
+AATCTG
+AATCTT
+AATGAA
+AATGAC
+AATGAG
+AATGAT
+AATGCA
+AATGCC
+AATGCG
+AATGCT
+AATGGA
+AATGGC
+AATGGG
+AATGGT
+AATGTA
+AATGTC
+AATGTG
+AATGTT
+AATTAA
+AATTAC
+AATTAG
+AATTAT
+AATTCA
+AATTCC
+AATTCG
+AATTCT
+AATTGA
+AATTGC
+AATTGG
+AATTGT
+AATTTA
+AATTTC
+AATTTG
+AATTTT
+ACAAAA
+ACAAAC
+ACAAAG
+ACAAAT
+ACAACA
+ACAACC
+ACAACG
+ACAACT
+ACAAGA
+ACAAGC
+ACAAGG
+ACAAGT
+ACAATA
+ACAATC
+ACAATG
+ACAATT
+ACACAA
+ACACAC
+ACACAG
+ACACAT
+ACACCA
+ACACCC
+ACACCG
+ACACCT
+ACACGA
+ACACGC
+ACACGG
+ACACGT
+ACACTA
+ACACTC
+ACACTG
+ACACTT
+ACAGAA
+ACAGAC
+ACAGAG
+ACAGAT
+ACAGCA
+ACAGCC
+ACAGCG
+ACAGCT
+ACAGGA
+ACAGGC
+ACAGGG
+ACAGGT
+ACAGTA
+ACAGTC
+ACAGTG
+ACAGTT
+ACATAA
+ACATAC
+ACATAG
+ACATAT
+ACATCA
+ACATCC
+ACATCG
+ACATCT
+ACATGA
+ACATGC
+ACATGG
+ACATGT
+ACATTA
+ACATTC
+ACATTG
+ACATTT
+ACCAAA
+ACCAAC
+ACCAAG
+ACCAAT
+ACCACA
+ACCACC
+ACCACG
+ACCACT
+ACCAGA
+ACCAGC
+ACCAGG
+ACCAGT
+ACCATA
+ACCATC
+ACCATG
+ACCATT
+ACCCAA
+ACCCAC
+ACCCAG
+ACCCAT
+ACCCCA
+ACCCCC
+ACCCCG
+ACCCCT
+ACCCGA
+ACCCGC
+ACCCGG
+ACCCGT
+ACCCTA
+ACCCTC
+ACCCTG
+ACCCTT
+ACCGAA
+ACCGAC
+ACCGAG
+ACCGAT
+ACCGCA
+ACCGCC
+ACCGCG
+ACCGCT
+ACCGGA
+ACCGGC
+ACCGGG
+ACCGGT
+ACCGTA
+ACCGTC
+ACCGTG
+ACCGTT
+ACCTAA
+ACCTAC
+ACCTAG
+ACCTAT
+ACCTCA
+ACCTCC
+ACCTCG
+ACCTCT
+ACCTGA
+ACCTGC
+ACCTGG
+ACCTGT
+ACCTTA
+ACCTTC
+ACCTTG
+ACCTTT
+ACGAAA
+ACGAAC
+ACGAAG
+ACGAAT
+ACGACA
+ACGACC
+ACGACG
+ACGACT
+ACGAGA
+ACGAGC
+ACGAGG
+ACGAGT
+ACGATA
+ACGATC
+ACGATG
+ACGATT
+ACGCAA
+ACGCAC
+ACGCAG
+ACGCAT
+ACGCCA
+ACGCCC
+ACGCCG
+ACGCCT
+ACGCGA
+ACGCGC
+ACGCGG
+ACGCGT
+ACGCTA
+ACGCTC
+ACGCTG
+ACGCTT
+ACGGAA
+ACGGAC
+ACGGAG
+ACGGAT
+ACGGCA
+ACGGCC
+ACGGCG
+ACGGCT
+ACGGGA
+ACGGGC
+ACGGGG
+ACGGGT
+ACGGTA
+ACGGTC
+ACGGTG
+ACGGTT
+ACGTAA
+ACGTAC
+ACGTAG
+ACGTAT
+ACGTCA
+ACGTCC
+ACGTCG
+ACGTCT
+ACGTGA
+ACGTGC
+ACGTGG
+ACGTGT
+ACGTTA
+ACGTTC
+ACGTTG
+ACGTTT
+ACTAAA
+ACTAAC
+ACTAAG
+ACTAAT
+ACTACA
+ACTACC
+ACTACG
+ACTACT
+ACTAGA
+ACTAGC
+ACTAGG
+ACTAGT
+ACTATA
+ACTATC
+ACTATG
+ACTATT
+ACTCAA
+ACTCAC
+ACTCAG
+ACTCAT
+ACTCCA
+ACTCCC
+ACTCCG
+ACTCCT
+ACTCGA
+ACTCGC
+ACTCGG
+ACTCGT
+ACTCTA
+ACTCTC
+ACTCTG
+ACTCTT
+ACTGAA
+ACTGAC
+ACTGAG
+ACTGAT
+ACTGCA
+ACTGCC
+ACTGCG
+ACTGCT
+ACTGGA
+ACTGGC
+ACTGGG
+ACTGGT
+ACTGTA
+ACTGTC
+ACTGTG
+ACTGTT
+ACTTAA
+ACTTAC
+ACTTAG
+ACTTAT
+ACTTCA
+ACTTCC
+ACTTCG
+ACTTCT
+ACTTGA
+ACTTGC
+ACTTGG
+ACTTGT
+ACTTTA
+ACTTTC
+ACTTTG
+ACTTTT
+AGAAAA
+AGAAAC
+AGAAAG
+AGAAAT
+AGAACA
+AGAACC
+AGAACG
+AGAACT
+AGAAGA
+AGAAGC
+AGAAGG
+AGAAGT
+AGAATA
+AGAATC
+AGAATG
+AGAATT
+AGACAA
+AGACAC
+AGACAG
+AGACAT
+AGACCA
+AGACCC
+AGACCG
+AGACCT
+AGACGA
+AGACGC
+AGACGG
+AGACGT
+AGACTA
+AGACTC
+AGACTG
+AGACTT
+AGAGAA
+AGAGAC
+AGAGAG
+AGAGAT
+AGAGCA
+AGAGCC
+AGAGCG
+AGAGCT
+AGAGGA
+AGAGGC
+AGAGGG
+AGAGGT
+AGAGTA
+AGAGTC
+AGAGTG
+AGAGTT
+AGATAA
+AGATAC
+AGATAG
+AGATAT
+AGATCA
+AGATCC
+AGATCG
+AGATCT
+AGATGA
+AGATGC
+AGATGG
+AGATGT
+AGATTA
+AGATTC
+AGATTG
+AGATTT
+AGCAAA
+AGCAAC
+AGCAAG
+AGCAAT
+AGCACA
+AGCACC
+AGCACG
+AGCACT
+AGCAGA
+AGCAGC
+AGCAGG
+AGCAGT
+AGCATA
+AGCATC
+AGCATG
+AGCATT
+AGCCAA
+AGCCAC
+AGCCAG
+AGCCAT
+AGCCCA
+AGCCCC
+AGCCCG
+AGCCCT
+AGCCGA
+AGCCGC
+AGCCGG
+AGCCGT
+AGCCTA
+AGCCTC
+AGCCTG
+AGCCTT
+AGCGAA
+AGCGAC
+AGCGAG
+AGCGAT
+AGCGCA
+AGCGCC
+AGCGCG
+AGCGCT
+AGCGGA
+AGCGGC
+AGCGGG
+AGCGGT
+AGCGTA
+AGCGTC
+AGCGTG
+AGCGTT
+AGCTAA
+AGCTAC
+AGCTAG
+AGCTAT
+AGCTCA
+AGCTCC
+AGCTCG
+AGCTCT
+AGCTGA
+AGCTGC
+AGCTGG
+AGCTGT
+AGCTTA
+AGCTTC
+AGCTTG
+AGCTTT
+AGGAAA
+AGGAAC
+AGGAAG
+AGGAAT
+AGGACA
+AGGACC
+AGGACG
+AGGACT
+AGGAGA
+AGGAGC
+AGGAGG
+AGGAGT
+AGGATA
+AGGATC
+AGGATG
+AGGATT
+AGGCAA
+AGGCAC
+AGGCAG
+AGGCAT
+AGGCCA
+AGGCCC
+AGGCCG
+AGGCCT
+AGGCGA
+AGGCGC
+AGGCGG
+AGGCGT
+AGGCTA
+AGGCTC
+AGGCTG
+AGGCTT
+AGGGAA
+AGGGAC
+AGGGAG
+AGGGAT
+AGGGCA
+AGGGCC
+AGGGCG
+AGGGCT
+AGGGGA
+AGGGGC
+AGGGGG
+AGGGGT
+AGGGTA
+AGGGTC
+AGGGTG
+AGGGTT
+AGGTAA
+AGGTAC
+AGGTAG
+AGGTAT
+AGGTCA
+AGGTCC
+AGGTCG
+AGGTCT
+AGGTGA
+AGGTGC
+AGGTGG
+AGGTGT
+AGGTTA
+AGGTTC
+AGGTTG
+AGGTTT
+AGTAAA
+AGTAAC
+AGTAAG
+AGTAAT
+AGTACA
+AGTACC
+AGTACG
+AGTACT
+AGTAGA
+AGTAGC
+AGTAGG
+AGTAGT
+AGTATA
+AGTATC
+AGTATG
+AGTATT
+AGTCAA
+AGTCAC
+AGTCAG
+AGTCAT
+AGTCCA
+AGTCCC
+AGTCCG
+AGTCCT
+AGTCGA
+AGTCGC
+AGTCGG
+AGTCGT
+AGTCTA
+AGTCTC
+AGTCTG
+AGTCTT
+AGTGAA
+AGTGAC
+AGTGAG
+AGTGAT
+AGTGCA
+AGTGCC
+AGTGCG
+AGTGCT
+AGTGGA
+AGTGGC
+AGTGGG
+AGTGGT
+AGTGTA
+AGTGTC
+AGTGTG
+AGTGTT
+AGTTAA
+AGTTAC
+AGTTAG
+AGTTAT
+AGTTCA
+AGTTCC
+AGTTCG
+AGTTCT
+AGTTGA
+AGTTGC
+AGTTGG
+AGTTGT
+AGTTTA
+AGTTTC
+AGTTTG
+AGTTTT
+ATAAAA
+ATAAAC
+ATAAAG
+ATAAAT
+ATAACA
+ATAACC
+ATAACG
+ATAACT
+ATAAGA
+ATAAGC
+ATAAGG
+ATAAGT
+ATAATA
+ATAATC
+ATAATG
+ATAATT
+ATACAA
+ATACAC
+ATACAG
+ATACAT
+ATACCA
+ATACCC
+ATACCG
+ATACCT
+ATACGA
+ATACGC
+ATACGG
+ATACGT
+ATACTA
+ATACTC
+ATACTG
+ATACTT
+ATAGAA
+ATAGAC
+ATAGAG
+ATAGAT
+ATAGCA
+ATAGCC
+ATAGCG
+ATAGCT
+ATAGGA
+ATAGGC
+ATAGGG
+ATAGGT
+ATAGTA
+ATAGTC
+ATAGTG
+ATAGTT
+ATATAA
+ATATAC
+ATATAG
+ATATAT
+ATATCA
+ATATCC
+ATATCG
+ATATCT
+ATATGA
+ATATGC
+ATATGG
+ATATGT
+ATATTA
+ATATTC
+ATATTG
+ATATTT
+ATCAAA
+ATCAAC
+ATCAAG
+ATCAAT
+ATCACA
+ATCACC
+ATCACG
+ATCACT
+ATCAGA
+ATCAGC
+ATCAGG
+ATCAGT
+ATCATA
+ATCATC
+ATCATG
+ATCATT
+ATCCAA
+ATCCAC
+ATCCAG
+ATCCAT
+ATCCCA
+ATCCCC
+ATCCCG
+ATCCCT
+ATCCGA
+ATCCGC
+ATCCGG
+ATCCGT
+ATCCTA
+ATCCTC
+ATCCTG
+ATCCTT
+ATCGAA
+ATCGAC
+ATCGAG
+ATCGAT
+ATCGCA
+ATCGCC
+ATCGCG
+ATCGCT
+ATCGGA
+ATCGGC
+ATCGGG
+ATCGGT
+ATCGTA
+ATCGTC
+ATCGTG
+ATCGTT
+ATCTAA
+ATCTAC
+ATCTAG
+ATCTAT
+ATCTCA
+ATCTCC
+ATCTCG
+ATCTCT
+ATCTGA
+ATCTGC
+ATCTGG
+ATCTGT
+ATCTTA
+ATCTTC
+ATCTTG
+ATCTTT
+ATGAAA
+ATGAAC
+ATGAAG
+ATGAAT
+ATGACA
+ATGACC
+ATGACG
+ATGACT
+ATGAGA
+ATGAGC
+ATGAGG
+ATGAGT
+ATGATA
+ATGATC
+ATGATG
+ATGATT
+ATGCAA
+ATGCAC
+ATGCAG
+ATGCAT
+ATGCCA
+ATGCCC
+ATGCCG
+ATGCCT
+ATGCGA
+ATGCGC
+ATGCGG
+ATGCGT
+ATGCTA
+ATGCTC
+ATGCTG
+ATGCTT
+ATGGAA
+ATGGAC
+ATGGAG
+ATGGAT
+ATGGCA
+ATGGCC
+ATGGCG
+ATGGCT
+ATGGGA
+ATGGGC
+ATGGGG
+ATGGGT
+ATGGTA
+ATGGTC
+ATGGTG
+ATGGTT
+ATGTAA
+ATGTAC
+ATGTAG
+ATGTAT
+ATGTCA
+ATGTCC
+ATGTCG
+ATGTCT
+ATGTGA
+ATGTGC
+ATGTGG
+ATGTGT
+ATGTTA
+ATGTTC
+ATGTTG
+ATGTTT
+ATTAAA
+ATTAAC
+ATTAAG
+ATTAAT
+ATTACA
+ATTACC
+ATTACG
+ATTACT
+ATTAGA
+ATTAGC
+ATTAGG
+ATTAGT
+ATTATA
+ATTATC
+ATTATG
+ATTATT
+ATTCAA
+ATTCAC
+ATTCAG
+ATTCAT
+ATTCCA
+ATTCCC
+ATTCCG
+ATTCCT
+ATTCGA
+ATTCGC
+ATTCGG
+ATTCGT
+ATTCTA
+ATTCTC
+ATTCTG
+ATTCTT
+ATTGAA
+ATTGAC
+ATTGAG
+ATTGAT
+ATTGCA
+ATTGCC
+ATTGCG
+ATTGCT
+ATTGGA
+ATTGGC
+ATTGGG
+ATTGGT
+ATTGTA
+ATTGTC
+ATTGTG
+ATTGTT
+ATTTAA
+ATTTAC
+ATTTAG
+ATTTAT
+ATTTCA
+ATTTCC
+ATTTCG
+ATTTCT
+ATTTGA
+ATTTGC
+ATTTGG
+ATTTGT
+ATTTTA
+ATTTTC
+ATTTTG
+ATTTTT
+CAAAAA
+CAAAAC
+CAAAAG
+CAAAAT
+CAAACA
+CAAACC
+CAAACG
+CAAACT
+CAAAGA
+CAAAGC
+CAAAGG
+CAAAGT
+CAAATA
+CAAATC
+CAAATG
+CAAATT
+CAACAA
+CAACAC
+CAACAG
+CAACAT
+CAACCA
+CAACCC
+CAACCG
+CAACCT
+CAACGA
+CAACGC
+CAACGG
+CAACGT
+CAACTA
+CAACTC
+CAACTG
+CAACTT
+CAAGAA
+CAAGAC
+CAAGAG
+CAAGAT
+CAAGCA
+CAAGCC
+CAAGCG
+CAAGCT
+CAAGGA
+CAAGGC
+CAAGGG
+CAAGGT
+CAAGTA
+CAAGTC
+CAAGTG
+CAAGTT
+CAATAA
+CAATAC
+CAATAG
+CAATAT
+CAATCA
+CAATCC
+CAATCG
+CAATCT
+CAATGA
+CAATGC
+CAATGG
+CAATGT
+CAATTA
+CAATTC
+CAATTG
+CAATTT
+CACAAA
+CACAAC
+CACAAG
+CACAAT
+CACACA
+CACACC
+CACACG
+CACACT
+CACAGA
+CACAGC
+CACAGG
+CACAGT
+CACATA
+CACATC
+CACATG
+CACATT
+CACCAA
+CACCAC
+CACCAG
+CACCAT
+CACCCA
+CACCCC
+CACCCG
+CACCCT
+CACCGA
+CACCGC
+CACCGG
+CACCGT
+CACCTA
+CACCTC
+CACCTG
+CACCTT
+CACGAA
+CACGAC
+CACGAG
+CACGAT
+CACGCA
+CACGCC
+CACGCG
+CACGCT
+CACGGA
+CACGGC
+CACGGG
+CACGGT
+CACGTA
+CACGTC
+CACGTG
+CACGTT
+CACTAA
+CACTAC
+CACTAG
+CACTAT
+CACTCA
+CACTCC
+CACTCG
+CACTCT
+CACTGA
+CACTGC
+CACTGG
+CACTGT
+CACTTA
+CACTTC
+CACTTG
+CACTTT
+CAGAAA
+CAGAAC
+CAGAAG
+CAGAAT
+CAGACA
+CAGACC
+CAGACG
+CAGACT
+CAGAGA
+CAGAGC
+CAGAGG
+CAGAGT
+CAGATA
+CAGATC
+CAGATG
+CAGATT
+CAGCAA
+CAGCAC
+CAGCAG
+CAGCAT
+CAGCCA
+CAGCCC
+CAGCCG
+CAGCCT
+CAGCGA
+CAGCGC
+CAGCGG
+CAGCGT
+CAGCTA
+CAGCTC
+CAGCTG
+CAGCTT
+CAGGAA
+CAGGAC
+CAGGAG
+CAGGAT
+CAGGCA
+CAGGCC
+CAGGCG
+CAGGCT
+CAGGGA
+CAGGGC
+CAGGGG
+CAGGGT
+CAGGTA
+CAGGTC
+CAGGTG
+CAGGTT
+CAGTAA
+CAGTAC
+CAGTAG
+CAGTAT
+CAGTCA
+CAGTCC
+CAGTCG
+CAGTCT
+CAGTGA
+CAGTGC
+CAGTGG
+CAGTGT
+CAGTTA
+CAGTTC
+CAGTTG
+CAGTTT
+CATAAA
+CATAAC
+CATAAG
+CATAAT
+CATACA
+CATACC
+CATACG
+CATACT
+CATAGA
+CATAGC
+CATAGG
+CATAGT
+CATATA
+CATATC
+CATATG
+CATATT
+CATCAA
+CATCAC
+CATCAG
+CATCAT
+CATCCA
+CATCCC
+CATCCG
+CATCCT
+CATCGA
+CATCGC
+CATCGG
+CATCGT
+CATCTA
+CATCTC
+CATCTG
+CATCTT
+CATGAA
+CATGAC
+CATGAG
+CATGAT
+CATGCA
+CATGCC
+CATGCG
+CATGCT
+CATGGA
+CATGGC
+CATGGG
+CATGGT
+CATGTA
+CATGTC
+CATGTG
+CATGTT
+CATTAA
+CATTAC
+CATTAG
+CATTAT
+CATTCA
+CATTCC
+CATTCG
+CATTCT
+CATTGA
+CATTGC
+CATTGG
+CATTGT
+CATTTA
+CATTTC
+CATTTG
+CATTTT
+CCAAAA
+CCAAAC
+CCAAAG
+CCAAAT
+CCAACA
+CCAACC
+CCAACG
+CCAACT
+CCAAGA
+CCAAGC
+CCAAGG
+CCAAGT
+CCAATA
+CCAATC
+CCAATG
+CCAATT
+CCACAA
+CCACAC
+CCACAG
+CCACAT
+CCACCA
+CCACCC
+CCACCG
+CCACCT
+CCACGA
+CCACGC
+CCACGG
+CCACGT
+CCACTA
+CCACTC
+CCACTG
+CCACTT
+CCAGAA
+CCAGAC
+CCAGAG
+CCAGAT
+CCAGCA
+CCAGCC
+CCAGCG
+CCAGCT
+CCAGGA
+CCAGGC
+CCAGGG
+CCAGGT
+CCAGTA
+CCAGTC
+CCAGTG
+CCAGTT
+CCATAA
+CCATAC
+CCATAG
+CCATAT
+CCATCA
+CCATCC
+CCATCG
+CCATCT
+CCATGA
+CCATGC
+CCATGG
+CCATGT
+CCATTA
+CCATTC
+CCATTG
+CCATTT
+CCCAAA
+CCCAAC
+CCCAAG
+CCCAAT
+CCCACA
+CCCACC
+CCCACG
+CCCACT
+CCCAGA
+CCCAGC
+CCCAGG
+CCCAGT
+CCCATA
+CCCATC
+CCCATG
+CCCATT
+CCCCAA
+CCCCAC
+CCCCAG
+CCCCAT
+CCCCCA
+CCCCCC
+CCCCCG
+CCCCCT
+CCCCGA
+CCCCGC
+CCCCGG
+CCCCGT
+CCCCTA
+CCCCTC
+CCCCTG
+CCCCTT
+CCCGAA
+CCCGAC
+CCCGAG
+CCCGAT
+CCCGCA
+CCCGCC
+CCCGCG
+CCCGCT
+CCCGGA
+CCCGGC
+CCCGGG
+CCCGGT
+CCCGTA
+CCCGTC
+CCCGTG
+CCCGTT
+CCCTAA
+CCCTAC
+CCCTAG
+CCCTAT
+CCCTCA
+CCCTCC
+CCCTCG
+CCCTCT
+CCCTGA
+CCCTGC
+CCCTGG
+CCCTGT
+CCCTTA
+CCCTTC
+CCCTTG
+CCCTTT
+CCGAAA
+CCGAAC
+CCGAAG
+CCGAAT
+CCGACA
+CCGACC
+CCGACG
+CCGACT
+CCGAGA
+CCGAGC
+CCGAGG
+CCGAGT
+CCGATA
+CCGATC
+CCGATG
+CCGATT
+CCGCAA
+CCGCAC
+CCGCAG
+CCGCAT
+CCGCCA
+CCGCCC
+CCGCCG
+CCGCCT
+CCGCGA
+CCGCGC
+CCGCGG
+CCGCGT
+CCGCTA
+CCGCTC
+CCGCTG
+CCGCTT
+CCGGAA
+CCGGAC
+CCGGAG
+CCGGAT
+CCGGCA
+CCGGCC
+CCGGCG
+CCGGCT
+CCGGGA
+CCGGGC
+CCGGGG
+CCGGGT
+CCGGTA
+CCGGTC
+CCGGTG
+CCGGTT
+CCGTAA
+CCGTAC
+CCGTAG
+CCGTAT
+CCGTCA
+CCGTCC
+CCGTCG
+CCGTCT
+CCGTGA
+CCGTGC
+CCGTGG
+CCGTGT
+CCGTTA
+CCGTTC
+CCGTTG
+CCGTTT
+CCTAAA
+CCTAAC
+CCTAAG
+CCTAAT
+CCTACA
+CCTACC
+CCTACG
+CCTACT
+CCTAGA
+CCTAGC
+CCTAGG
+CCTAGT
+CCTATA
+CCTATC
+CCTATG
+CCTATT
+CCTCAA
+CCTCAC
+CCTCAG
+CCTCAT
+CCTCCA
+CCTCCC
+CCTCCG
+CCTCCT
+CCTCGA
+CCTCGC
+CCTCGG
+CCTCGT
+CCTCTA
+CCTCTC
+CCTCTG
+CCTCTT
+CCTGAA
+CCTGAC
+CCTGAG
+CCTGAT
+CCTGCA
+CCTGCC
+CCTGCG
+CCTGCT
+CCTGGA
+CCTGGC
+CCTGGG
+CCTGGT
+CCTGTA
+CCTGTC
+CCTGTG
+CCTGTT
+CCTTAA
+CCTTAC
+CCTTAG
+CCTTAT
+CCTTCA
+CCTTCC
+CCTTCG
+CCTTCT
+CCTTGA
+CCTTGC
+CCTTGG
+CCTTGT
+CCTTTA
+CCTTTC
+CCTTTG
+CCTTTT
+CGAAAA
+CGAAAC
+CGAAAG
+CGAAAT
+CGAACA
+CGAACC
+CGAACG
+CGAACT
+CGAAGA
+CGAAGC
+CGAAGG
+CGAAGT
+CGAATA
+CGAATC
+CGAATG
+CGAATT
+CGACAA
+CGACAC
+CGACAG
+CGACAT
+CGACCA
+CGACCC
+CGACCG
+CGACCT
+CGACGA
+CGACGC
+CGACGG
+CGACGT
+CGACTA
+CGACTC
+CGACTG
+CGACTT
+CGAGAA
+CGAGAC
+CGAGAG
+CGAGAT
+CGAGCA
+CGAGCC
+CGAGCG
+CGAGCT
+CGAGGA
+CGAGGC
+CGAGGG
+CGAGGT
+CGAGTA
+CGAGTC
+CGAGTG
+CGAGTT
+CGATAA
+CGATAC
+CGATAG
+CGATAT
+CGATCA
+CGATCC
+CGATCG
+CGATCT
+CGATGA
+CGATGC
+CGATGG
+CGATGT
+CGATTA
+CGATTC
+CGATTG
+CGATTT
+CGCAAA
+CGCAAC
+CGCAAG
+CGCAAT
+CGCACA
+CGCACC
+CGCACG
+CGCACT
+CGCAGA
+CGCAGC
+CGCAGG
+CGCAGT
+CGCATA
+CGCATC
+CGCATG
+CGCATT
+CGCCAA
+CGCCAC
+CGCCAG
+CGCCAT
+CGCCCA
+CGCCCC
+CGCCCG
+CGCCCT
+CGCCGA
+CGCCGC
+CGCCGG
+CGCCGT
+CGCCTA
+CGCCTC
+CGCCTG
+CGCCTT
+CGCGAA
+CGCGAC
+CGCGAG
+CGCGAT
+CGCGCA
+CGCGCC
+CGCGCG
+CGCGCT
+CGCGGA
+CGCGGC
+CGCGGG
+CGCGGT
+CGCGTA
+CGCGTC
+CGCGTG
+CGCGTT
+CGCTAA
+CGCTAC
+CGCTAG
+CGCTAT
+CGCTCA
+CGCTCC
+CGCTCG
+CGCTCT
+CGCTGA
+CGCTGC
+CGCTGG
+CGCTGT
+CGCTTA
+CGCTTC
+CGCTTG
+CGCTTT
+CGGAAA
+CGGAAC
+CGGAAG
+CGGAAT
+CGGACA
+CGGACC
+CGGACG
+CGGACT
+CGGAGA
+CGGAGC
+CGGAGG
+CGGAGT
+CGGATA
+CGGATC
+CGGATG
+CGGATT
+CGGCAA
+CGGCAC
+CGGCAG
+CGGCAT
+CGGCCA
+CGGCCC
+CGGCCG
+CGGCCT
+CGGCGA
+CGGCGC
+CGGCGG
+CGGCGT
+CGGCTA
+CGGCTC
+CGGCTG
+CGGCTT
+CGGGAA
+CGGGAC
+CGGGAG
+CGGGAT
+CGGGCA
+CGGGCC
+CGGGCG
+CGGGCT
+CGGGGA
+CGGGGC
+CGGGGG
+CGGGGT
+CGGGTA
+CGGGTC
+CGGGTG
+CGGGTT
+CGGTAA
+CGGTAC
+CGGTAG
+CGGTAT
+CGGTCA
+CGGTCC
+CGGTCG
+CGGTCT
+CGGTGA
+CGGTGC
+CGGTGG
+CGGTGT
+CGGTTA
+CGGTTC
+CGGTTG
+CGGTTT
+CGTAAA
+CGTAAC
+CGTAAG
+CGTAAT
+CGTACA
+CGTACC
+CGTACG
+CGTACT
+CGTAGA
+CGTAGC
+CGTAGG
+CGTAGT
+CGTATA
+CGTATC
+CGTATG
+CGTATT
+CGTCAA
+CGTCAC
+CGTCAG
+CGTCAT
+CGTCCA
+CGTCCC
+CGTCCG
+CGTCCT
+CGTCGA
+CGTCGC
+CGTCGG
+CGTCGT
+CGTCTA
+CGTCTC
+CGTCTG
+CGTCTT
+CGTGAA
+CGTGAC
+CGTGAG
+CGTGAT
+CGTGCA
+CGTGCC
+CGTGCG
+CGTGCT
+CGTGGA
+CGTGGC
+CGTGGG
+CGTGGT
+CGTGTA
+CGTGTC
+CGTGTG
+CGTGTT
+CGTTAA
+CGTTAC
+CGTTAG
+CGTTAT
+CGTTCA
+CGTTCC
+CGTTCG
+CGTTCT
+CGTTGA
+CGTTGC
+CGTTGG
+CGTTGT
+CGTTTA
+CGTTTC
+CGTTTG
+CGTTTT
+CTAAAA
+CTAAAC
+CTAAAG
+CTAAAT
+CTAACA
+CTAACC
+CTAACG
+CTAACT
+CTAAGA
+CTAAGC
+CTAAGG
+CTAAGT
+CTAATA
+CTAATC
+CTAATG
+CTAATT
+CTACAA
+CTACAC
+CTACAG
+CTACAT
+CTACCA
+CTACCC
+CTACCG
+CTACCT
+CTACGA
+CTACGC
+CTACGG
+CTACGT
+CTACTA
+CTACTC
+CTACTG
+CTACTT
+CTAGAA
+CTAGAC
+CTAGAG
+CTAGAT
+CTAGCA
+CTAGCC
+CTAGCG
+CTAGCT
+CTAGGA
+CTAGGC
+CTAGGG
+CTAGGT
+CTAGTA
+CTAGTC
+CTAGTG
+CTAGTT
+CTATAA
+CTATAC
+CTATAG
+CTATAT
+CTATCA
+CTATCC
+CTATCG
+CTATCT
+CTATGA
+CTATGC
+CTATGG
+CTATGT
+CTATTA
+CTATTC
+CTATTG
+CTATTT
+CTCAAA
+CTCAAC
+CTCAAG
+CTCAAT
+CTCACA
+CTCACC
+CTCACG
+CTCACT
+CTCAGA
+CTCAGC
+CTCAGG
+CTCAGT
+CTCATA
+CTCATC
+CTCATG
+CTCATT
+CTCCAA
+CTCCAC
+CTCCAG
+CTCCAT
+CTCCCA
+CTCCCC
+CTCCCG
+CTCCCT
+CTCCGA
+CTCCGC
+CTCCGG
+CTCCGT
+CTCCTA
+CTCCTC
+CTCCTG
+CTCCTT
+CTCGAA
+CTCGAC
+CTCGAG
+CTCGAT
+CTCGCA
+CTCGCC
+CTCGCG
+CTCGCT
+CTCGGA
+CTCGGC
+CTCGGG
+CTCGGT
+CTCGTA
+CTCGTC
+CTCGTG
+CTCGTT
+CTCTAA
+CTCTAC
+CTCTAG
+CTCTAT
+CTCTCA
+CTCTCC
+CTCTCG
+CTCTCT
+CTCTGA
+CTCTGC
+CTCTGG
+CTCTGT
+CTCTTA
+CTCTTC
+CTCTTG
+CTCTTT
+CTGAAA
+CTGAAC
+CTGAAG
+CTGAAT
+CTGACA
+CTGACC
+CTGACG
+CTGACT
+CTGAGA
+CTGAGC
+CTGAGG
+CTGAGT
+CTGATA
+CTGATC
+CTGATG
+CTGATT
+CTGCAA
+CTGCAC
+CTGCAG
+CTGCAT
+CTGCCA
+CTGCCC
+CTGCCG
+CTGCCT
+CTGCGA
+CTGCGC
+CTGCGG
+CTGCGT
+CTGCTA
+CTGCTC
+CTGCTG
+CTGCTT
+CTGGAA
+CTGGAC
+CTGGAG
+CTGGAT
+CTGGCA
+CTGGCC
+CTGGCG
+CTGGCT
+CTGGGA
+CTGGGC
+CTGGGG
+CTGGGT
+CTGGTA
+CTGGTC
+CTGGTG
+CTGGTT
+CTGTAA
+CTGTAC
+CTGTAG
+CTGTAT
+CTGTCA
+CTGTCC
+CTGTCG
+CTGTCT
+CTGTGA
+CTGTGC
+CTGTGG
+CTGTGT
+CTGTTA
+CTGTTC
+CTGTTG
+CTGTTT
+CTTAAA
+CTTAAC
+CTTAAG
+CTTAAT
+CTTACA
+CTTACC
+CTTACG
+CTTACT
+CTTAGA
+CTTAGC
+CTTAGG
+CTTAGT
+CTTATA
+CTTATC
+CTTATG
+CTTATT
+CTTCAA
+CTTCAC
+CTTCAG
+CTTCAT
+CTTCCA
+CTTCCC
+CTTCCG
+CTTCCT
+CTTCGA
+CTTCGC
+CTTCGG
+CTTCGT
+CTTCTA
+CTTCTC
+CTTCTG
+CTTCTT
+CTTGAA
+CTTGAC
+CTTGAG
+CTTGAT
+CTTGCA
+CTTGCC
+CTTGCG
+CTTGCT
+CTTGGA
+CTTGGC
+CTTGGG
+CTTGGT
+CTTGTA
+CTTGTC
+CTTGTG
+CTTGTT
+CTTTAA
+CTTTAC
+CTTTAG
+CTTTAT
+CTTTCA
+CTTTCC
+CTTTCG
+CTTTCT
+CTTTGA
+CTTTGC
+CTTTGG
+CTTTGT
+CTTTTA
+CTTTTC
+CTTTTG
+CTTTTT
+GAAAAA
+GAAAAC
+GAAAAG
+GAAAAT
+GAAACA
+GAAACC
+GAAACG
+GAAACT
+GAAAGA
+GAAAGC
+GAAAGG
+GAAAGT
+GAAATA
+GAAATC
+GAAATG
+GAAATT
+GAACAA
+GAACAC
+GAACAG
+GAACAT
+GAACCA
+GAACCC
+GAACCG
+GAACCT
+GAACGA
+GAACGC
+GAACGG
+GAACGT
+GAACTA
+GAACTC
+GAACTG
+GAACTT
+GAAGAA
+GAAGAC
+GAAGAG
+GAAGAT
+GAAGCA
+GAAGCC
+GAAGCG
+GAAGCT
+GAAGGA
+GAAGGC
+GAAGGG
+GAAGGT
+GAAGTA
+GAAGTC
+GAAGTG
+GAAGTT
+GAATAA
+GAATAC
+GAATAG
+GAATAT
+GAATCA
+GAATCC
+GAATCG
+GAATCT
+GAATGA
+GAATGC
+GAATGG
+GAATGT
+GAATTA
+GAATTC
+GAATTG
+GAATTT
+GACAAA
+GACAAC
+GACAAG
+GACAAT
+GACACA
+GACACC
+GACACG
+GACACT
+GACAGA
+GACAGC
+GACAGG
+GACAGT
+GACATA
+GACATC
+GACATG
+GACATT
+GACCAA
+GACCAC
+GACCAG
+GACCAT
+GACCCA
+GACCCC
+GACCCG
+GACCCT
+GACCGA
+GACCGC
+GACCGG
+GACCGT
+GACCTA
+GACCTC
+GACCTG
+GACCTT
+GACGAA
+GACGAC
+GACGAG
+GACGAT
+GACGCA
+GACGCC
+GACGCG
+GACGCT
+GACGGA
+GACGGC
+GACGGG
+GACGGT
+GACGTA
+GACGTC
+GACGTG
+GACGTT
+GACTAA
+GACTAC
+GACTAG
+GACTAT
+GACTCA
+GACTCC
+GACTCG
+GACTCT
+GACTGA
+GACTGC
+GACTGG
+GACTGT
+GACTTA
+GACTTC
+GACTTG
+GACTTT
+GAGAAA
+GAGAAC
+GAGAAG
+GAGAAT
+GAGACA
+GAGACC
+GAGACG
+GAGACT
+GAGAGA
+GAGAGC
+GAGAGG
+GAGAGT
+GAGATA
+GAGATC
+GAGATG
+GAGATT
+GAGCAA
+GAGCAC
+GAGCAG
+GAGCAT
+GAGCCA
+GAGCCC
+GAGCCG
+GAGCCT
+GAGCGA
+GAGCGC
+GAGCGG
+GAGCGT
+GAGCTA
+GAGCTC
+GAGCTG
+GAGCTT
+GAGGAA
+GAGGAC
+GAGGAG
+GAGGAT
+GAGGCA
+GAGGCC
+GAGGCG
+GAGGCT
+GAGGGA
+GAGGGC
+GAGGGG
+GAGGGT
+GAGGTA
+GAGGTC
+GAGGTG
+GAGGTT
+GAGTAA
+GAGTAC
+GAGTAG
+GAGTAT
+GAGTCA
+GAGTCC
+GAGTCG
+GAGTCT
+GAGTGA
+GAGTGC
+GAGTGG
+GAGTGT
+GAGTTA
+GAGTTC
+GAGTTG
+GAGTTT
+GATAAA
+GATAAC
+GATAAG
+GATAAT
+GATACA
+GATACC
+GATACG
+GATACT
+GATAGA
+GATAGC
+GATAGG
+GATAGT
+GATATA
+GATATC
+GATATG
+GATATT
+GATCAA
+GATCAC
+GATCAG
+GATCAT
+GATCCA
+GATCCC
+GATCCG
+GATCCT
+GATCGA
+GATCGC
+GATCGG
+GATCGT
+GATCTA
+GATCTC
+GATCTG
+GATCTT
+GATGAA
+GATGAC
+GATGAG
+GATGAT
+GATGCA
+GATGCC
+GATGCG
+GATGCT
+GATGGA
+GATGGC
+GATGGG
+GATGGT
+GATGTA
+GATGTC
+GATGTG
+GATGTT
+GATTAA
+GATTAC
+GATTAG
+GATTAT
+GATTCA
+GATTCC
+GATTCG
+GATTCT
+GATTGA
+GATTGC
+GATTGG
+GATTGT
+GATTTA
+GATTTC
+GATTTG
+GATTTT
+GCAAAA
+GCAAAC
+GCAAAG
+GCAAAT
+GCAACA
+GCAACC
+GCAACG
+GCAACT
+GCAAGA
+GCAAGC
+GCAAGG
+GCAAGT
+GCAATA
+GCAATC
+GCAATG
+GCAATT
+GCACAA
+GCACAC
+GCACAG
+GCACAT
+GCACCA
+GCACCC
+GCACCG
+GCACCT
+GCACGA
+GCACGC
+GCACGG
+GCACGT
+GCACTA
+GCACTC
+GCACTG
+GCACTT
+GCAGAA
+GCAGAC
+GCAGAG
+GCAGAT
+GCAGCA
+GCAGCC
+GCAGCG
+GCAGCT
+GCAGGA
+GCAGGC
+GCAGGG
+GCAGGT
+GCAGTA
+GCAGTC
+GCAGTG
+GCAGTT
+GCATAA
+GCATAC
+GCATAG
+GCATAT
+GCATCA
+GCATCC
+GCATCG
+GCATCT
+GCATGA
+GCATGC
+GCATGG
+GCATGT
+GCATTA
+GCATTC
+GCATTG
+GCATTT
+GCCAAA
+GCCAAC
+GCCAAG
+GCCAAT
+GCCACA
+GCCACC
+GCCACG
+GCCACT
+GCCAGA
+GCCAGC
+GCCAGG
+GCCAGT
+GCCATA
+GCCATC
+GCCATG
+GCCATT
+GCCCAA
+GCCCAC
+GCCCAG
+GCCCAT
+GCCCCA
+GCCCCC
+GCCCCG
+GCCCCT
+GCCCGA
+GCCCGC
+GCCCGG
+GCCCGT
+GCCCTA
+GCCCTC
+GCCCTG
+GCCCTT
+GCCGAA
+GCCGAC
+GCCGAG
+GCCGAT
+GCCGCA
+GCCGCC
+GCCGCG
+GCCGCT
+GCCGGA
+GCCGGC
+GCCGGG
+GCCGGT
+GCCGTA
+GCCGTC
+GCCGTG
+GCCGTT
+GCCTAA
+GCCTAC
+GCCTAG
+GCCTAT
+GCCTCA
+GCCTCC
+GCCTCG
+GCCTCT
+GCCTGA
+GCCTGC
+GCCTGG
+GCCTGT
+GCCTTA
+GCCTTC
+GCCTTG
+GCCTTT
+GCGAAA
+GCGAAC
+GCGAAG
+GCGAAT
+GCGACA
+GCGACC
+GCGACG
+GCGACT
+GCGAGA
+GCGAGC
+GCGAGG
+GCGAGT
+GCGATA
+GCGATC
+GCGATG
+GCGATT
+GCGCAA
+GCGCAC
+GCGCAG
+GCGCAT
+GCGCCA
+GCGCCC
+GCGCCG
+GCGCCT
+GCGCGA
+GCGCGC
+GCGCGG
+GCGCGT
+GCGCTA
+GCGCTC
+GCGCTG
+GCGCTT
+GCGGAA
+GCGGAC
+GCGGAG
+GCGGAT
+GCGGCA
+GCGGCC
+GCGGCG
+GCGGCT
+GCGGGA
+GCGGGC
+GCGGGG
+GCGGGT
+GCGGTA
+GCGGTC
+GCGGTG
+GCGGTT
+GCGTAA
+GCGTAC
+GCGTAG
+GCGTAT
+GCGTCA
+GCGTCC
+GCGTCG
+GCGTCT
+GCGTGA
+GCGTGC
+GCGTGG
+GCGTGT
+GCGTTA
+GCGTTC
+GCGTTG
+GCGTTT
+GCTAAA
+GCTAAC
+GCTAAG
+GCTAAT
+GCTACA
+GCTACC
+GCTACG
+GCTACT
+GCTAGA
+GCTAGC
+GCTAGG
+GCTAGT
+GCTATA
+GCTATC
+GCTATG
+GCTATT
+GCTCAA
+GCTCAC
+GCTCAG
+GCTCAT
+GCTCCA
+GCTCCC
+GCTCCG
+GCTCCT
+GCTCGA
+GCTCGC
+GCTCGG
+GCTCGT
+GCTCTA
+GCTCTC
+GCTCTG
+GCTCTT
+GCTGAA
+GCTGAC
+GCTGAG
+GCTGAT
+GCTGCA
+GCTGCC
+GCTGCG
+GCTGCT
+GCTGGA
+GCTGGC
+GCTGGG
+GCTGGT
+GCTGTA
+GCTGTC
+GCTGTG
+GCTGTT
+GCTTAA
+GCTTAC
+GCTTAG
+GCTTAT
+GCTTCA
+GCTTCC
+GCTTCG
+GCTTCT
+GCTTGA
+GCTTGC
+GCTTGG
+GCTTGT
+GCTTTA
+GCTTTC
+GCTTTG
+GCTTTT
+GGAAAA
+GGAAAC
+GGAAAG
+GGAAAT
+GGAACA
+GGAACC
+GGAACG
+GGAACT
+GGAAGA
+GGAAGC
+GGAAGG
+GGAAGT
+GGAATA
+GGAATC
+GGAATG
+GGAATT
+GGACAA
+GGACAC
+GGACAG
+GGACAT
+GGACCA
+GGACCC
+GGACCG
+GGACCT
+GGACGA
+GGACGC
+GGACGG
+GGACGT
+GGACTA
+GGACTC
+GGACTG
+GGACTT
+GGAGAA
+GGAGAC
+GGAGAG
+GGAGAT
+GGAGCA
+GGAGCC
+GGAGCG
+GGAGCT
+GGAGGA
+GGAGGC
+GGAGGG
+GGAGGT
+GGAGTA
+GGAGTC
+GGAGTG
+GGAGTT
+GGATAA
+GGATAC
+GGATAG
+GGATAT
+GGATCA
+GGATCC
+GGATCG
+GGATCT
+GGATGA
+GGATGC
+GGATGG
+GGATGT
+GGATTA
+GGATTC
+GGATTG
+GGATTT
+GGCAAA
+GGCAAC
+GGCAAG
+GGCAAT
+GGCACA
+GGCACC
+GGCACG
+GGCACT
+GGCAGA
+GGCAGC
+GGCAGG
+GGCAGT
+GGCATA
+GGCATC
+GGCATG
+GGCATT
+GGCCAA
+GGCCAC
+GGCCAG
+GGCCAT
+GGCCCA
+GGCCCC
+GGCCCG
+GGCCCT
+GGCCGA
+GGCCGC
+GGCCGG
+GGCCGT
+GGCCTA
+GGCCTC
+GGCCTG
+GGCCTT
+GGCGAA
+GGCGAC
+GGCGAG
+GGCGAT
+GGCGCA
+GGCGCC
+GGCGCG
+GGCGCT
+GGCGGA
+GGCGGC
+GGCGGG
+GGCGGT
+GGCGTA
+GGCGTC
+GGCGTG
+GGCGTT
+GGCTAA
+GGCTAC
+GGCTAG
+GGCTAT
+GGCTCA
+GGCTCC
+GGCTCG
+GGCTCT
+GGCTGA
+GGCTGC
+GGCTGG
+GGCTGT
+GGCTTA
+GGCTTC
+GGCTTG
+GGCTTT
+GGGAAA
+GGGAAC
+GGGAAG
+GGGAAT
+GGGACA
+GGGACC
+GGGACG
+GGGACT
+GGGAGA
+GGGAGC
+GGGAGG
+GGGAGT
+GGGATA
+GGGATC
+GGGATG
+GGGATT
+GGGCAA
+GGGCAC
+GGGCAG
+GGGCAT
+GGGCCA
+GGGCCC
+GGGCCG
+GGGCCT
+GGGCGA
+GGGCGC
+GGGCGG
+GGGCGT
+GGGCTA
+GGGCTC
+GGGCTG
+GGGCTT
+GGGGAA
+GGGGAC
+GGGGAG
+GGGGAT
+GGGGCA
+GGGGCC
+GGGGCG
+GGGGCT
+GGGGGA
+GGGGGC
+GGGGGG
+GGGGGT
+GGGGTA
+GGGGTC
+GGGGTG
+GGGGTT
+GGGTAA
+GGGTAC
+GGGTAG
+GGGTAT
+GGGTCA
+GGGTCC
+GGGTCG
+GGGTCT
+GGGTGA
+GGGTGC
+GGGTGG
+GGGTGT
+GGGTTA
+GGGTTC
+GGGTTG
+GGGTTT
+GGTAAA
+GGTAAC
+GGTAAG
+GGTAAT
+GGTACA
+GGTACC
+GGTACG
+GGTACT
+GGTAGA
+GGTAGC
+GGTAGG
+GGTAGT
+GGTATA
+GGTATC
+GGTATG
+GGTATT
+GGTCAA
+GGTCAC
+GGTCAG
+GGTCAT
+GGTCCA
+GGTCCC
+GGTCCG
+GGTCCT
+GGTCGA
+GGTCGC
+GGTCGG
+GGTCGT
+GGTCTA
+GGTCTC
+GGTCTG
+GGTCTT
+GGTGAA
+GGTGAC
+GGTGAG
+GGTGAT
+GGTGCA
+GGTGCC
+GGTGCG
+GGTGCT
+GGTGGA
+GGTGGC
+GGTGGG
+GGTGGT
+GGTGTA
+GGTGTC
+GGTGTG
+GGTGTT
+GGTTAA
+GGTTAC
+GGTTAG
+GGTTAT
+GGTTCA
+GGTTCC
+GGTTCG
+GGTTCT
+GGTTGA
+GGTTGC
+GGTTGG
+GGTTGT
+GGTTTA
+GGTTTC
+GGTTTG
+GGTTTT
+GTAAAA
+GTAAAC
+GTAAAG
+GTAAAT
+GTAACA
+GTAACC
+GTAACG
+GTAACT
+GTAAGA
+GTAAGC
+GTAAGG
+GTAAGT
+GTAATA
+GTAATC
+GTAATG
+GTAATT
+GTACAA
+GTACAC
+GTACAG
+GTACAT
+GTACCA
+GTACCC
+GTACCG
+GTACCT
+GTACGA
+GTACGC
+GTACGG
+GTACGT
+GTACTA
+GTACTC
+GTACTG
+GTACTT
+GTAGAA
+GTAGAC
+GTAGAG
+GTAGAT
+GTAGCA
+GTAGCC
+GTAGCG
+GTAGCT
+GTAGGA
+GTAGGC
+GTAGGG
+GTAGGT
+GTAGTA
+GTAGTC
+GTAGTG
+GTAGTT
+GTATAA
+GTATAC
+GTATAG
+GTATAT
+GTATCA
+GTATCC
+GTATCG
+GTATCT
+GTATGA
+GTATGC
+GTATGG
+GTATGT
+GTATTA
+GTATTC
+GTATTG
+GTATTT
+GTCAAA
+GTCAAC
+GTCAAG
+GTCAAT
+GTCACA
+GTCACC
+GTCACG
+GTCACT
+GTCAGA
+GTCAGC
+GTCAGG
+GTCAGT
+GTCATA
+GTCATC
+GTCATG
+GTCATT
+GTCCAA
+GTCCAC
+GTCCAG
+GTCCAT
+GTCCCA
+GTCCCC
+GTCCCG
+GTCCCT
+GTCCGA
+GTCCGC
+GTCCGG
+GTCCGT
+GTCCTA
+GTCCTC
+GTCCTG
+GTCCTT
+GTCGAA
+GTCGAC
+GTCGAG
+GTCGAT
+GTCGCA
+GTCGCC
+GTCGCG
+GTCGCT
+GTCGGA
+GTCGGC
+GTCGGG
+GTCGGT
+GTCGTA
+GTCGTC
+GTCGTG
+GTCGTT
+GTCTAA
+GTCTAC
+GTCTAG
+GTCTAT
+GTCTCA
+GTCTCC
+GTCTCG
+GTCTCT
+GTCTGA
+GTCTGC
+GTCTGG
+GTCTGT
+GTCTTA
+GTCTTC
+GTCTTG
+GTCTTT
+GTGAAA
+GTGAAC
+GTGAAG
+GTGAAT
+GTGACA
+GTGACC
+GTGACG
+GTGACT
+GTGAGA
+GTGAGC
+GTGAGG
+GTGAGT
+GTGATA
+GTGATC
+GTGATG
+GTGATT
+GTGCAA
+GTGCAC
+GTGCAG
+GTGCAT
+GTGCCA
+GTGCCC
+GTGCCG
+GTGCCT
+GTGCGA
+GTGCGC
+GTGCGG
+GTGCGT
+GTGCTA
+GTGCTC
+GTGCTG
+GTGCTT
+GTGGAA
+GTGGAC
+GTGGAG
+GTGGAT
+GTGGCA
+GTGGCC
+GTGGCG
+GTGGCT
+GTGGGA
+GTGGGC
+GTGGGG
+GTGGGT
+GTGGTA
+GTGGTC
+GTGGTG
+GTGGTT
+GTGTAA
+GTGTAC
+GTGTAG
+GTGTAT
+GTGTCA
+GTGTCC
+GTGTCG
+GTGTCT
+GTGTGA
+GTGTGC
+GTGTGG
+GTGTGT
+GTGTTA
+GTGTTC
+GTGTTG
+GTGTTT
+GTTAAA
+GTTAAC
+GTTAAG
+GTTAAT
+GTTACA
+GTTACC
+GTTACG
+GTTACT
+GTTAGA
+GTTAGC
+GTTAGG
+GTTAGT
+GTTATA
+GTTATC
+GTTATG
+GTTATT
+GTTCAA
+GTTCAC
+GTTCAG
+GTTCAT
+GTTCCA
+GTTCCC
+GTTCCG
+GTTCCT
+GTTCGA
+GTTCGC
+GTTCGG
+GTTCGT
+GTTCTA
+GTTCTC
+GTTCTG
+GTTCTT
+GTTGAA
+GTTGAC
+GTTGAG
+GTTGAT
+GTTGCA
+GTTGCC
+GTTGCG
+GTTGCT
+GTTGGA
+GTTGGC
+GTTGGG
+GTTGGT
+GTTGTA
+GTTGTC
+GTTGTG
+GTTGTT
+GTTTAA
+GTTTAC
+GTTTAG
+GTTTAT
+GTTTCA
+GTTTCC
+GTTTCG
+GTTTCT
+GTTTGA
+GTTTGC
+GTTTGG
+GTTTGT
+GTTTTA
+GTTTTC
+GTTTTG
+GTTTTT
+TAAAAA
+TAAAAC
+TAAAAG
+TAAAAT
+TAAACA
+TAAACC
+TAAACG
+TAAACT
+TAAAGA
+TAAAGC
+TAAAGG
+TAAAGT
+TAAATA
+TAAATC
+TAAATG
+TAAATT
+TAACAA
+TAACAC
+TAACAG
+TAACAT
+TAACCA
+TAACCC
+TAACCG
+TAACCT
+TAACGA
+TAACGC
+TAACGG
+TAACGT
+TAACTA
+TAACTC
+TAACTG
+TAACTT
+TAAGAA
+TAAGAC
+TAAGAG
+TAAGAT
+TAAGCA
+TAAGCC
+TAAGCG
+TAAGCT
+TAAGGA
+TAAGGC
+TAAGGG
+TAAGGT
+TAAGTA
+TAAGTC
+TAAGTG
+TAAGTT
+TAATAA
+TAATAC
+TAATAG
+TAATAT
+TAATCA
+TAATCC
+TAATCG
+TAATCT
+TAATGA
+TAATGC
+TAATGG
+TAATGT
+TAATTA
+TAATTC
+TAATTG
+TAATTT
+TACAAA
+TACAAC
+TACAAG
+TACAAT
+TACACA
+TACACC
+TACACG
+TACACT
+TACAGA
+TACAGC
+TACAGG
+TACAGT
+TACATA
+TACATC
+TACATG
+TACATT
+TACCAA
+TACCAC
+TACCAG
+TACCAT
+TACCCA
+TACCCC
+TACCCG
+TACCCT
+TACCGA
+TACCGC
+TACCGG
+TACCGT
+TACCTA
+TACCTC
+TACCTG
+TACCTT
+TACGAA
+TACGAC
+TACGAG
+TACGAT
+TACGCA
+TACGCC
+TACGCG
+TACGCT
+TACGGA
+TACGGC
+TACGGG
+TACGGT
+TACGTA
+TACGTC
+TACGTG
+TACGTT
+TACTAA
+TACTAC
+TACTAG
+TACTAT
+TACTCA
+TACTCC
+TACTCG
+TACTCT
+TACTGA
+TACTGC
+TACTGG
+TACTGT
+TACTTA
+TACTTC
+TACTTG
+TACTTT
+TAGAAA
+TAGAAC
+TAGAAG
+TAGAAT
+TAGACA
+TAGACC
+TAGACG
+TAGACT
+TAGAGA
+TAGAGC
+TAGAGG
+TAGAGT
+TAGATA
+TAGATC
+TAGATG
+TAGATT
+TAGCAA
+TAGCAC
+TAGCAG
+TAGCAT
+TAGCCA
+TAGCCC
+TAGCCG
+TAGCCT
+TAGCGA
+TAGCGC
+TAGCGG
+TAGCGT
+TAGCTA
+TAGCTC
+TAGCTG
+TAGCTT
+TAGGAA
+TAGGAC
+TAGGAG
+TAGGAT
+TAGGCA
+TAGGCC
+TAGGCG
+TAGGCT
+TAGGGA
+TAGGGC
+TAGGGG
+TAGGGT
+TAGGTA
+TAGGTC
+TAGGTG
+TAGGTT
+TAGTAA
+TAGTAC
+TAGTAG
+TAGTAT
+TAGTCA
+TAGTCC
+TAGTCG
+TAGTCT
+TAGTGA
+TAGTGC
+TAGTGG
+TAGTGT
+TAGTTA
+TAGTTC
+TAGTTG
+TAGTTT
+TATAAA
+TATAAC
+TATAAG
+TATAAT
+TATACA
+TATACC
+TATACG
+TATACT
+TATAGA
+TATAGC
+TATAGG
+TATAGT
+TATATA
+TATATC
+TATATG
+TATATT
+TATCAA
+TATCAC
+TATCAG
+TATCAT
+TATCCA
+TATCCC
+TATCCG
+TATCCT
+TATCGA
+TATCGC
+TATCGG
+TATCGT
+TATCTA
+TATCTC
+TATCTG
+TATCTT
+TATGAA
+TATGAC
+TATGAG
+TATGAT
+TATGCA
+TATGCC
+TATGCG
+TATGCT
+TATGGA
+TATGGC
+TATGGG
+TATGGT
+TATGTA
+TATGTC
+TATGTG
+TATGTT
+TATTAA
+TATTAC
+TATTAG
+TATTAT
+TATTCA
+TATTCC
+TATTCG
+TATTCT
+TATTGA
+TATTGC
+TATTGG
+TATTGT
+TATTTA
+TATTTC
+TATTTG
+TATTTT
+TCAAAA
+TCAAAC
+TCAAAG
+TCAAAT
+TCAACA
+TCAACC
+TCAACG
+TCAACT
+TCAAGA
+TCAAGC
+TCAAGG
+TCAAGT
+TCAATA
+TCAATC
+TCAATG
+TCAATT
+TCACAA
+TCACAC
+TCACAG
+TCACAT
+TCACCA
+TCACCC
+TCACCG
+TCACCT
+TCACGA
+TCACGC
+TCACGG
+TCACGT
+TCACTA
+TCACTC
+TCACTG
+TCACTT
+TCAGAA
+TCAGAC
+TCAGAG
+TCAGAT
+TCAGCA
+TCAGCC
+TCAGCG
+TCAGCT
+TCAGGA
+TCAGGC
+TCAGGG
+TCAGGT
+TCAGTA
+TCAGTC
+TCAGTG
+TCAGTT
+TCATAA
+TCATAC
+TCATAG
+TCATAT
+TCATCA
+TCATCC
+TCATCG
+TCATCT
+TCATGA
+TCATGC
+TCATGG
+TCATGT
+TCATTA
+TCATTC
+TCATTG
+TCATTT
+TCCAAA
+TCCAAC
+TCCAAG
+TCCAAT
+TCCACA
+TCCACC
+TCCACG
+TCCACT
+TCCAGA
+TCCAGC
+TCCAGG
+TCCAGT
+TCCATA
+TCCATC
+TCCATG
+TCCATT
+TCCCAA
+TCCCAC
+TCCCAG
+TCCCAT
+TCCCCA
+TCCCCC
+TCCCCG
+TCCCCT
+TCCCGA
+TCCCGC
+TCCCGG
+TCCCGT
+TCCCTA
+TCCCTC
+TCCCTG
+TCCCTT
+TCCGAA
+TCCGAC
+TCCGAG
+TCCGAT
+TCCGCA
+TCCGCC
+TCCGCG
+TCCGCT
+TCCGGA
+TCCGGC
+TCCGGG
+TCCGGT
+TCCGTA
+TCCGTC
+TCCGTG
+TCCGTT
+TCCTAA
+TCCTAC
+TCCTAG
+TCCTAT
+TCCTCA
+TCCTCC
+TCCTCG
+TCCTCT
+TCCTGA
+TCCTGC
+TCCTGG
+TCCTGT
+TCCTTA
+TCCTTC
+TCCTTG
+TCCTTT
+TCGAAA
+TCGAAC
+TCGAAG
+TCGAAT
+TCGACA
+TCGACC
+TCGACG
+TCGACT
+TCGAGA
+TCGAGC
+TCGAGG
+TCGAGT
+TCGATA
+TCGATC
+TCGATG
+TCGATT
+TCGCAA
+TCGCAC
+TCGCAG
+TCGCAT
+TCGCCA
+TCGCCC
+TCGCCG
+TCGCCT
+TCGCGA
+TCGCGC
+TCGCGG
+TCGCGT
+TCGCTA
+TCGCTC
+TCGCTG
+TCGCTT
+TCGGAA
+TCGGAC
+TCGGAG
+TCGGAT
+TCGGCA
+TCGGCC
+TCGGCG
+TCGGCT
+TCGGGA
+TCGGGC
+TCGGGG
+TCGGGT
+TCGGTA
+TCGGTC
+TCGGTG
+TCGGTT
+TCGTAA
+TCGTAC
+TCGTAG
+TCGTAT
+TCGTCA
+TCGTCC
+TCGTCG
+TCGTCT
+TCGTGA
+TCGTGC
+TCGTGG
+TCGTGT
+TCGTTA
+TCGTTC
+TCGTTG
+TCGTTT
+TCTAAA
+TCTAAC
+TCTAAG
+TCTAAT
+TCTACA
+TCTACC
+TCTACG
+TCTACT
+TCTAGA
+TCTAGC
+TCTAGG
+TCTAGT
+TCTATA
+TCTATC
+TCTATG
+TCTATT
+TCTCAA
+TCTCAC
+TCTCAG
+TCTCAT
+TCTCCA
+TCTCCC
+TCTCCG
+TCTCCT
+TCTCGA
+TCTCGC
+TCTCGG
+TCTCGT
+TCTCTA
+TCTCTC
+TCTCTG
+TCTCTT
+TCTGAA
+TCTGAC
+TCTGAG
+TCTGAT
+TCTGCA
+TCTGCC
+TCTGCG
+TCTGCT
+TCTGGA
+TCTGGC
+TCTGGG
+TCTGGT
+TCTGTA
+TCTGTC
+TCTGTG
+TCTGTT
+TCTTAA
+TCTTAC
+TCTTAG
+TCTTAT
+TCTTCA
+TCTTCC
+TCTTCG
+TCTTCT
+TCTTGA
+TCTTGC
+TCTTGG
+TCTTGT
+TCTTTA
+TCTTTC
+TCTTTG
+TCTTTT
+TGAAAA
+TGAAAC
+TGAAAG
+TGAAAT
+TGAACA
+TGAACC
+TGAACG
+TGAACT
+TGAAGA
+TGAAGC
+TGAAGG
+TGAAGT
+TGAATA
+TGAATC
+TGAATG
+TGAATT
+TGACAA
+TGACAC
+TGACAG
+TGACAT
+TGACCA
+TGACCC
+TGACCG
+TGACCT
+TGACGA
+TGACGC
+TGACGG
+TGACGT
+TGACTA
+TGACTC
+TGACTG
+TGACTT
+TGAGAA
+TGAGAC
+TGAGAG
+TGAGAT
+TGAGCA
+TGAGCC
+TGAGCG
+TGAGCT
+TGAGGA
+TGAGGC
+TGAGGG
+TGAGGT
+TGAGTA
+TGAGTC
+TGAGTG
+TGAGTT
+TGATAA
+TGATAC
+TGATAG
+TGATAT
+TGATCA
+TGATCC
+TGATCG
+TGATCT
+TGATGA
+TGATGC
+TGATGG
+TGATGT
+TGATTA
+TGATTC
+TGATTG
+TGATTT
+TGCAAA
+TGCAAC
+TGCAAG
+TGCAAT
+TGCACA
+TGCACC
+TGCACG
+TGCACT
+TGCAGA
+TGCAGC
+TGCAGG
+TGCAGT
+TGCATA
+TGCATC
+TGCATG
+TGCATT
+TGCCAA
+TGCCAC
+TGCCAG
+TGCCAT
+TGCCCA
+TGCCCC
+TGCCCG
+TGCCCT
+TGCCGA
+TGCCGC
+TGCCGG
+TGCCGT
+TGCCTA
+TGCCTC
+TGCCTG
+TGCCTT
+TGCGAA
+TGCGAC
+TGCGAG
+TGCGAT
+TGCGCA
+TGCGCC
+TGCGCG
+TGCGCT
+TGCGGA
+TGCGGC
+TGCGGG
+TGCGGT
+TGCGTA
+TGCGTC
+TGCGTG
+TGCGTT
+TGCTAA
+TGCTAC
+TGCTAG
+TGCTAT
+TGCTCA
+TGCTCC
+TGCTCG
+TGCTCT
+TGCTGA
+TGCTGC
+TGCTGG
+TGCTGT
+TGCTTA
+TGCTTC
+TGCTTG
+TGCTTT
+TGGAAA
+TGGAAC
+TGGAAG
+TGGAAT
+TGGACA
+TGGACC
+TGGACG
+TGGACT
+TGGAGA
+TGGAGC
+TGGAGG
+TGGAGT
+TGGATA
+TGGATC
+TGGATG
+TGGATT
+TGGCAA
+TGGCAC
+TGGCAG
+TGGCAT
+TGGCCA
+TGGCCC
+TGGCCG
+TGGCCT
+TGGCGA
+TGGCGC
+TGGCGG
+TGGCGT
+TGGCTA
+TGGCTC
+TGGCTG
+TGGCTT
+TGGGAA
+TGGGAC
+TGGGAG
+TGGGAT
+TGGGCA
+TGGGCC
+TGGGCG
+TGGGCT
+TGGGGA
+TGGGGC
+TGGGGG
+TGGGGT
+TGGGTA
+TGGGTC
+TGGGTG
+TGGGTT
+TGGTAA
+TGGTAC
+TGGTAG
+TGGTAT
+TGGTCA
+TGGTCC
+TGGTCG
+TGGTCT
+TGGTGA
+TGGTGC
+TGGTGG
+TGGTGT
+TGGTTA
+TGGTTC
+TGGTTG
+TGGTTT
+TGTAAA
+TGTAAC
+TGTAAG
+TGTAAT
+TGTACA
+TGTACC
+TGTACG
+TGTACT
+TGTAGA
+TGTAGC
+TGTAGG
+TGTAGT
+TGTATA
+TGTATC
+TGTATG
+TGTATT
+TGTCAA
+TGTCAC
+TGTCAG
+TGTCAT
+TGTCCA
+TGTCCC
+TGTCCG
+TGTCCT
+TGTCGA
+TGTCGC
+TGTCGG
+TGTCGT
+TGTCTA
+TGTCTC
+TGTCTG
+TGTCTT
+TGTGAA
+TGTGAC
+TGTGAG
+TGTGAT
+TGTGCA
+TGTGCC
+TGTGCG
+TGTGCT
+TGTGGA
+TGTGGC
+TGTGGG
+TGTGGT
+TGTGTA
+TGTGTC
+TGTGTG
+TGTGTT
+TGTTAA
+TGTTAC
+TGTTAG
+TGTTAT
+TGTTCA
+TGTTCC
+TGTTCG
+TGTTCT
+TGTTGA
+TGTTGC
+TGTTGG
+TGTTGT
+TGTTTA
+TGTTTC
+TGTTTG
+TGTTTT
+TTAAAA
+TTAAAC
+TTAAAG
+TTAAAT
+TTAACA
+TTAACC
+TTAACG
+TTAACT
+TTAAGA
+TTAAGC
+TTAAGG
+TTAAGT
+TTAATA
+TTAATC
+TTAATG
+TTAATT
+TTACAA
+TTACAC
+TTACAG
+TTACAT
+TTACCA
+TTACCC
+TTACCG
+TTACCT
+TTACGA
+TTACGC
+TTACGG
+TTACGT
+TTACTA
+TTACTC
+TTACTG
+TTACTT
+TTAGAA
+TTAGAC
+TTAGAG
+TTAGAT
+TTAGCA
+TTAGCC
+TTAGCG
+TTAGCT
+TTAGGA
+TTAGGC
+TTAGGG
+TTAGGT
+TTAGTA
+TTAGTC
+TTAGTG
+TTAGTT
+TTATAA
+TTATAC
+TTATAG
+TTATAT
+TTATCA
+TTATCC
+TTATCG
+TTATCT
+TTATGA
+TTATGC
+TTATGG
+TTATGT
+TTATTA
+TTATTC
+TTATTG
+TTATTT
+TTCAAA
+TTCAAC
+TTCAAG
+TTCAAT
+TTCACA
+TTCACC
+TTCACG
+TTCACT
+TTCAGA
+TTCAGC
+TTCAGG
+TTCAGT
+TTCATA
+TTCATC
+TTCATG
+TTCATT
+TTCCAA
+TTCCAC
+TTCCAG
+TTCCAT
+TTCCCA
+TTCCCC
+TTCCCG
+TTCCCT
+TTCCGA
+TTCCGC
+TTCCGG
+TTCCGT
+TTCCTA
+TTCCTC
+TTCCTG
+TTCCTT
+TTCGAA
+TTCGAC
+TTCGAG
+TTCGAT
+TTCGCA
+TTCGCC
+TTCGCG
+TTCGCT
+TTCGGA
+TTCGGC
+TTCGGG
+TTCGGT
+TTCGTA
+TTCGTC
+TTCGTG
+TTCGTT
+TTCTAA
+TTCTAC
+TTCTAG
+TTCTAT
+TTCTCA
+TTCTCC
+TTCTCG
+TTCTCT
+TTCTGA
+TTCTGC
+TTCTGG
+TTCTGT
+TTCTTA
+TTCTTC
+TTCTTG
+TTCTTT
+TTGAAA
+TTGAAC
+TTGAAG
+TTGAAT
+TTGACA
+TTGACC
+TTGACG
+TTGACT
+TTGAGA
+TTGAGC
+TTGAGG
+TTGAGT
+TTGATA
+TTGATC
+TTGATG
+TTGATT
+TTGCAA
+TTGCAC
+TTGCAG
+TTGCAT
+TTGCCA
+TTGCCC
+TTGCCG
+TTGCCT
+TTGCGA
+TTGCGC
+TTGCGG
+TTGCGT
+TTGCTA
+TTGCTC
+TTGCTG
+TTGCTT
+TTGGAA
+TTGGAC
+TTGGAG
+TTGGAT
+TTGGCA
+TTGGCC
+TTGGCG
+TTGGCT
+TTGGGA
+TTGGGC
+TTGGGG
+TTGGGT
+TTGGTA
+TTGGTC
+TTGGTG
+TTGGTT
+TTGTAA
+TTGTAC
+TTGTAG
+TTGTAT
+TTGTCA
+TTGTCC
+TTGTCG
+TTGTCT
+TTGTGA
+TTGTGC
+TTGTGG
+TTGTGT
+TTGTTA
+TTGTTC
+TTGTTG
+TTGTTT
+TTTAAA
+TTTAAC
+TTTAAG
+TTTAAT
+TTTACA
+TTTACC
+TTTACG
+TTTACT
+TTTAGA
+TTTAGC
+TTTAGG
+TTTAGT
+TTTATA
+TTTATC
+TTTATG
+TTTATT
+TTTCAA
+TTTCAC
+TTTCAG
+TTTCAT
+TTTCCA
+TTTCCC
+TTTCCG
+TTTCCT
+TTTCGA
+TTTCGC
+TTTCGG
+TTTCGT
+TTTCTA
+TTTCTC
+TTTCTG
+TTTCTT
+TTTGAA
+TTTGAC
+TTTGAG
+TTTGAT
+TTTGCA
+TTTGCC
+TTTGCG
+TTTGCT
+TTTGGA
+TTTGGC
+TTTGGG
+TTTGGT
+TTTGTA
+TTTGTC
+TTTGTG
+TTTGTT
+TTTTAA
+TTTTAC
+TTTTAG
+TTTTAT
+TTTTCA
+TTTTCC
+TTTTCG
+TTTTCT
+TTTTGA
+TTTTGC
+TTTTGG
+TTTTGT
+TTTTTA
+TTTTTC
+TTTTTG
+TTTTTT
+NNNNNN