ChIP-seq workflow

[jxtx]

Create a workflow for peak calling VISION ChIP-seq datasets that (as closely as possible) matches the ENCODE best practices pipelines. ENCODE pipelines are here but they are completely tied to the DNANexus platform so we cannot use.

• [x] Map with bwa mem
• [x] Mark duplicates
• [x] Remove unmapped reads
• [x] Remove blacklisted regions
• [ ] Calcualate NRF, PBC1, PBC2
• [ ] Perform cross-correlation analysis
• [x] Peak call with MACS2
• [ ] Generate fold-over-control and signal tracks with MACS2.
  • [ ] Tools for bedGraphToBigWig and macs bdgcmp
• [ ] Support narrow and broad peaks
  • [ ] ENCODE uses SPP for "narrow" peaks
• [x] Do replicates independently
• [x] Also do pooled
• [x] Perform IDR analysis (using pooled peaks as oracle)

[jxtx]

Another potential change / improvement may be to change how output filenames are done. I was renaming everything at the end, but it may make sense to just name things better along the way. CWL seems to like filenames to be meaningful.