Can a ChIP-seq .bedgraph file (four columns) be used as input for 'labels' in Catchitt about jstacs HOT 4 CLOSED

This should work, in general.

I just created a MWE based on our bed line:

File peaks.bed has one line:
chr2L 9180 9200 0.095
(just as you posted, but tab-separated)

File chr.fai has one line:
chr2L 10000 0 0 0

With those, I called the labels tool:

java -jar Catchitt-0.1.3.jar labels c=peaks.bed r=peaks.bed f=chr.fai b=50 rw=50

As output, I get a file Labels.tsv.gz with mostly "U" labels as you reported. However, there is also one "S" label as would be expected from the input and arguments of the call to labels

chr2L	9050	U
chr2L	9100	U
chr2L	9150	S
chr2L	9200	U
chr2L	9250	U

because only the 50 bp bin from 9150 to 9200 overlaps with the peak from the bed file.

Could you please check that our output does indeed not contain any labels different from "U"?

Besides, you should make sure that these are indeed peaks from a ChIP-seq experiment (or this input at least makes sense for your use case).
A peak width of 20 bp seems to be rather narrow.

from jstacs.

The .bedgraph file is from chip-seq data (description: 'Bedgraph files are 20 bp bins of reads from both replicates, scaled to reads-per-million of reads overlapping peak region' from https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSM3474126)

so there are no 'B' or 'S' labels in my 'label' output of catchitt for sure.

my .fai file has chromosome labels:
>NC_004354.4 Drosophila melanogaster chromosome 2L

while my chip seq bedgraph file looks like this:
chr2L 9180 9200 0.095
chr2L 9200 9220 0.19
chr2L 9220 9240 0.19
chr2L 9240 9260 0.19
...

is it possible the discrepancy in labeling (chromosome 2L vs chr2L) is the issue?

from jstacs.

You .fai file looks more like the header of a FastA file, especially with the line starting with ">".

And yes, the names of the sequences in the .fai must exactly match the sequence identifiers from the bed file.
(Same holds true for accessibility mapping files, etc. in later steps of Catchitt.)

from jstacs.

sorry ! that was the fasta file, the .fai file has labels like this "NC_004354.4" and NOT labels like in my chip file. face palm thank you

from jstacs.