MethylBERT documentation

The figure was generated using [biorender](https://www.biorender.com/)

BERT model to classify read-level DNA methylation data into tumour/normal and perform tumour deconvolution. MethylBERT is implemented using pytorch and transformers 🤗.

Publication

MethylBERT paper is now online on Nature Communications

MethylBERT enables read-level DNA methylation pattern identification and tumour deconvolution using a Transformer-based model

Yunhee Jeong, Clarissa Gerhäuser, Guido Sauter, Thorsten Schlomm, Karl Rohr and Pavlo Lutsik

Contents

• Data Preparation
  • Input requirements
  • Simulated read-level methylomes
• Using MethylBERT as a Python library
  • 01. Preprocessing for MethylBERT fine-tuning training data
  • 02. Preprocessing bulk data
  • 03. Fine-tuning MethylBERT model
  • 04. Tumour deconvolution
• Using MethylBERT as a command line interface (CLI)
  • Coming soon…