metadata

dataset_info:
  features:
    - name: text
      dtype: string
    - name: relation
      dtype: string
    - name: h
      struct:
        - name: id
          dtype: int64
        - name: name
          dtype: string
        - name: pos
          sequence: int64
    - name: t
      struct:
        - name: id
          dtype: string
        - name: name
          dtype: string
        - name: pos
          sequence: int64
  splits:
    - name: train
      num_bytes: 54491244
      num_examples: 178264
    - name: validation
      num_bytes: 6118764
      num_examples: 20193
    - name: test
      num_bytes: 6168865
      num_examples: 20516
  download_size: 35878376
  dataset_size: 66778873
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*
task_categories:
  - text-classification
language:
  - en
tags:
  - biology
  - relation-classification
  - medical
  - relation-extraction
  - gene
  - disease
  - gda
pretty_name: TBGA
size_categories:
  - 100K<n<1M

Dataset Card for TBGA

Dataset Description

Repository: https://github.com/GDAMining/gda-extraction
Paper: TBGA: a large‑scale Gene‑Disease Association dataset for Biomedical RelationExtraction

Dataset Summary

TBGA is a comprehensive dataset created for the purpose of Gene-Disease Association (GDA) extraction, generated from over 700,000 publications. It features more than 200,000 instances and 100,000 unique gene-disease pairs. Each instance in the dataset includes the specific sentence from which the GDA was extracted, the extracted GDA itself, and detailed information about the gene-disease pair involved. This dataset was semi-automatically annotated by Marchesin and Silvello using data sourced from the DisGeNET database, which houses one of the most extensive collections of genes and variants associated with human diseases. The dataset follows the OpenNRE format and contains the following relations:

{"NA": 0, "therapeutic": 1, "biomarker": 2, "genomic_alterations": 3}

Languages

The language in the dataset is English.

Dataset Structure

Dataset Instances

An example of 'train' looks as follows:

{
  "text": "A monocyte chemoattractant protein-1 gene polymorphism is associated with occult ischemia in a high-risk asymptomatic population.",
  "relation": "NA",
  "h": {
    "id": 6347,
    "name": "CCL2",
    "pos": [2, 34]
  },
  "t": {
    "id": "C0231221",
    "name": "Asymptomatic",
    "pos": [105, 12]
  }
}

Data Fields

text: the text of this example, a string feature.
h: the gene entity
- id: NCBI Entrez ID associated with the gene entity, a string feature.
- pos: list consisting of starting position and length of the gene mention withintext, a list of int32 features.
- name: NCBI official gene symbol associated with the gene entity (not the text of the mention), a string feature.
t: the disease entity
- id: UMLS Concept Unique Identifier (CUI) associated with the disease entity, a string feature.
- pos: list consisting of starting position and length of the disease mention withintext, a list of int32 features.
- name: UMLS preferred term associated with the disease entity (not the text of the mention), a string feature.
relation: a class label associated with the given GDA.

Citation

BibTeX:

@article{marchesin-silvello-2022,
  title = "TBGA: a large-scale Gene-Disease Association dataset for Biomedical Relation Extraction",
  author = "S. Marchesin and G. Silvello",
  journal = "BMC Bioinformatics",
  year = "2022",
  url = "https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-022-04646-6",
  doi = "10.1186/s12859-022-04646-6",
  volume = "23",
  number = "1",
  pages = "111"
}

APA:

Marchesin, S., & Silvello, G. (2022). TBGA: A large-scale Gene-Disease Association dataset for Biomedical Relation Extraction. BMC Bioinformatics, 23(1), 111. https://doi.org/10.1186/s12859-022-04646-6

Dataset Card Authors

@phucdev