---
license: mit
task_categories:
- feature-extraction
- token-classification
- zero-shot-classification
language:
- en
tags:
- biology
- medical
size_categories:
- 100B<n<1T
---

# Mobius: Mixture-Of-Experts Transformer Model in Epigenetics of ME/CFS and Long COVID

[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)

## Dataset Overview

This dataset contains DNA methylation data from blood samples of individuals with Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS), Long COVID (LC), and healthy controls. The data is derived from Illumina HumanMethylation450 BeadChip and Illumina MethylationEPIC arrays and has been organized to facilitate research into epigenetic biomarkers for these conditions.

## Dataset Structure

```
root/
├── ME/                   # Myalgic Encephalomyelitis/Chronic Fatigue Syndrome samples
│   ├── GSE102504/        # GEO Series accession number
│   │   └── GSE102504_RAW/    # Raw data directory
│   │       └── *.idat    # Illumina IDAT files
│   ├── GSE111183/
│   │   └── GSE111183_RAW/
│   │       └── *.idat
│   ├── GSE112905/
│   │   └── GSE112905_RAW/
│   │       └── *.idat
│   ├── GSE153667/
│   │   └── GSE153667_RAW/
│   │       └── *.idat
│   ├── GSE156792/
│   │   └── GSE156792_RAW/
│   │       └── *.idat
│   ├── GSE166592/
│   │   └── GSE166592_RAW/
│   │       └── *.idat
│   ├── GSE59489/
│   │   └── GSE59489_RAW/
│   │       └── *.idat
│   └── GSE93266/
│       └── GSE93266_RAW/
│           └── *.idat
├── LC/                   # Long COVID samples
│   ├── GSE161678/
│   │   └── GSE161678_RAW/
│   │       └── *.idat
│   ├── GSE174818/
│   │   └── GSE174818_RAW/
│   │       └── *.idat
│   ├── GSE188573/
│   │   └── GSE188573_RAW/
│   │       └── *.idat
│   ├── GSE197152/
│   │   └── GSE197152_RAW/
│   │       └── *.idat
│   └── GSE210430/
│       └── GSE210430_RAW/
│           └── *.idat
└── controls/             # Healthy control samples
    ├── GSE112905/
    │   └── GSE112905_RAW/
    │       └── *.idat
    ├── GSE40279/
    │   └── GSE40279_RAW/
    │       └── *.idat
    └── GSE42861/
        └── GSE42861_RAW/
            └── *.idat
```

## Data Description

- **File Format**: `.idat` (Illumina's proprietary binary file format for BeadArray data)
- **Platforms**: Illumina HumanMethylation450 BeadChip and Illumina MethylationEPIC arrays
- **Sample Composition**: 252 ME/CFS patients, 252 Long COVID patients, and 252 healthy controls
- **Tissue Type**: Peripheral blood
- **Data Type**: DNA methylation beta values (representing the ratio of methylated probe intensity to total probe intensity)

## About the Conditions

### ME/CFS (Myalgic Encephalomyelitis/Chronic Fatigue Syndrome)
- Chronic multi-system illness affecting approximately 0.5-1% of the population worldwide
- Often triggered by viral illnesses
- Characterized by long-term fatigue, post-exertional malaise, cognitive impairment, and autonomic dysfunction
- Currently diagnosed by clinical exclusion due to lack of definitive biomarkers

### Long COVID (Post-Acute Sequelae of COVID-19)
- Refers to prolonged symptoms after acute SARS-CoV-2 infection
- Affects approximately 10-20% of COVID-19 survivors
- Shares some clinical features with ME/CFS
- Similarly lacks definitive biomarkers and is diagnosed primarily by clinical assessment

## Usage Example

Here's a simple example of how to load and process this data using the `minfi` package in R:

```r
library(minfi)
library(limma)
library(wateRmelon)

# Load raw data
baseDir <- "path/to/dataset/ME/GSE111183/GSE111183_RAW"
targets <- read.metharray.sheet(baseDir)
rgSet <- read.metharray.exp(targets=targets)

# Preprocess
mSet <- preprocessRaw(rgSet)
beta <- getBeta(mSet)

# Normalization
beta_norm <- BMIQ(beta)

# Visualization example
densityPlot(beta_norm, sampGroups=targets$Group)
```

## Citation Information

If you use this dataset in your research, please cite:

```
Acharya, P., & Jacoby, D. (2025). Epigenomic Transformer Pipeline for ME/CFS and Long COVID Classification. 
University of Victoria.
```

## Additional Resources

For more information about the Epigenomic Transformer Pipeline developed using this data, please visit our [GitHub repository](https:github.com/VerisimilitudeX/EpiMECoV).

## Acknowledgements

Data was sourced from the NCBI Gene Expression Omnibus (GEO) with the following accession numbers:
- ME/CFS data: GSE102504, GSE111183, GSE112905, GSE153667, GSE156792, GSE166592, GSE59489, GSE93266
- Long COVID data: GSE161678, GSE174818, GSE188573, GSE197152, GSE210430
- Control data: GSE112905, GSE40279, GSE42861

We thank all the researchers who contributed to generating and sharing this valuable data.