Check for overlapping DMRs across comparisons

Next steps:

  • determine if there are better (more appropriate) parameters to be using when calling DMRs (see section below). Then redo the bedtools closet on DMRfind output if different than current output.

Compare outputs of different DMRfind parameters

Ambient sample DMRfind result files to compare

  1. min-cov = 5; mc-max-dist = 25, min-cluster = 2
  2. min-cov = 5; mc-max-dist = 25, min-cluster = 3
  3. min-cov = 5; mc-max-dist = 50, min-cluster = 3
  4. merged allc; min-cov = 5; mc-max-dist = 25, min-cluster = 3

Bigwig files to load into IGV

Creating bigwig files:

Summary table of comparison:

parameter –min-cluster –mc-max-dist –min-num-dms –dmr-max-dist –min-cov  
descr min #samples/group #bp between sites where mCG counts can be summed min #DMS for DMR to be reported max distance signif. sites can be to be included in same DMR min #reads for DMS to be considered #DMRs called
  2 25 3 250 5 192
  3 25 3 250 5 22
  3 50 3 250 5 32
  merge all samples/group 25 3 250 5 716

IGV session: https://gannet.fish.washington.edu/metacarcinus/Pgenerosa/analyses/20190822/merge_allc/amb_compareDMRfindParams.xml

  • Files loaded:
  • RESULTS:
    • DMR.bed files are ordered as follows:
      1. –min-cluster 3, –mc-max 25
      2. –min-cluster 3, –mc-max 50
      3. –min-cluster 2, –mc-max 25
      4. all samples merged per group, –mc-max 25
    • interesting that some chromosomes have no DMRs when all samples are merged
  • Next Steps:
    • visualize with 5x cov. filtered bedgraphs to confirm DMRs are not false positives
    • create .bw files from merged allc files to see how .bw files from individual samples compared to merged sample files.
      • this is to mainly understand what is going on during merging. From what I understand, the allc files are combined and overlapping CpG counts are summed across samles