Differing numbers of interactions reported in seqmonk and washU output files

[Russell Hamilton]

There appear to be different numbers of interactions reported in the seqmonk and washU format files (taking into account the 2 line vs 1 line format). Is this the expected outcome or should the numbers of interactions match?

The table below shows the line counts for a selection of my output files. Seqmonk and ibed are the same, but washU is consistently off by a few thousand

                 washU seqmonk  seqmonk/2   ibed
data1           123142  254394  127197      127198
data2           137212  281394  140697      140698
data3           149581  305370  152685      152686
dataMerged      171783  356412  178206      178207

[Jonathan Cairns]

Hi Russell,

This is because bait to bait interactions can be tested either way round. Seqmonk supports asymmetric interaction scores, whereas WashU does not so the maximum is taken. See:

http://regulatorygenomicsgroup.org/resources/Chicago_vignette.html#washu_text-format-ends-with-washu_text.txt

Thanks,

Jonathan