EPCOT: A computational tool to use ATAC-seq to impute epigenome, transcriptome, and high-resolution chromatin contact maps

[ Code / Documentation ] [ Paper ] [ Google Colab Notebook ]

Instructions:

  • Use the first app to process ATAC-seq data by entering a Google Drive or Dropbox link of ATAC-seq BAM file. It is recommaneded to use ENCODE ATAC-seq pipeline to generate the BAM file.
  • Execute the model by uploading the processed ATAC-seq file to the “Run Model” interface in the second app.
  • Visualize the prediction results by uploading the predicted files from the “Run Model” interface to the “Visualize Prediction Results” interface.

We provide detailed instructions for running the apps on Google Colab notebook with free GPUs Open In Colab . You can find necessary Python scripts on our GitHub repository.


App 1: Process ATAC-seq data

Notice:

  • Downsample the BAM file (to around 30M alignments) before upload if the size is too large. This helps prevent potential app downtime.
  • If processing takes too long or if the app is unavailable, you can alternatively run Python scripts on your own server to obtain the processed file.
  • Open in Hugging Face Space


    App 2: Run models and visualize prediction results

    Upload the processed ATAC-seq file from App1 and select a specific region of interest to run the predictive model.

    Open in Hugging Face Space