bcbl-kaist/DeepLUCIA_prep

creating features for DeepLUCIA ( https://github.com/bcbl-kaist/DeepLUCIA )

DeepLUCIA_prep

creating features for DeepLUCIA ( https://github.com/bcbl-kaist/DeepLUCIA )

Input data

• ChromInfo file for chromosome sizes. ( retrieved from UCSC goldenpath )
• BED files for gaps. ( retrieved from UCSC goldenpath )
• Bed file for blacklist regions. ( retrieved from ENCODE Project )
• The genome sequence fasta file is used to generate genomic feature file
• histone ChIP-seq -log(p) signal track/ in bigWig format is used to generate epigenomic feature file.
  • The Roadmap Epigenomics provides epigenome of more than 100 human tissues
  • If you have your own data, you can follow the guide from Roadmap Epigenomics ( https://egg2.wustl.edu/roadmap/web_portal/processed_data.html )
• chromatin loops file in bedpe format.

Parameters

• The length of genomic region to test ( DeepLUCIA used 5kb for default )
• The length of overlap between regions ( DeepLUCIA used 0bp for default )
• The resolution of epigenomic feature ( DeepLUCIA used 25bp for default )

Output files

• putativa anchor annotations.
• genomic loop annotations.
• One-hot encoded genomic feature npy file
• 25bp-resolution epigenomic feature npy file.

Dependencies

• pybedtools>=0.8.0 ( https://github.com/daler/pybedtools )
• pysam>=0.15.2 ( https://github.com/pysam-developers/pysam )
• pyBigWig>=0.3.13 ( https://github.com/deeptools/pyBigWig )