jaccard index is running significantly more comparisons than needs to be performed

[slsevilla]

Description of the bug

Running champagne on 12 samples with three contrasts. Checked on the runlogs since the pipeline was still in progress and noticed that JACCARD_INDEX populated over 1800 jobs!

[7b/3bdb74] process > CHIPSEQ:CALL_PEAKS:JACCARD_INDEX (macs_broad H3K4me1 ... [ 92%] 1747 of 1892

This is because of the way the input is being iterated here:

CHAMPAGNE/subworkflows/local/peaks.nf

Line 121 in d06ce0f

ch_peaks

Right now the index is being run on every sample:peak_caller comparison possible. Instead a single matrix should be created for each sample, comparing peak_callers within a sample.