Wig Files to Interval Files

Question

[gracie]

Hi

I am new in ChipSeq analysis and I am working on an online dataset. There is only wig files generated from MACS peak calling. However I need interval files with start site, end site, FDR etc. Is there anyway to convert wig files to interval files? Or any other way to obtain genomic intervals from those wig files?

Thanks

[GenoMax]

How about wig2bed from bedops?

If the files are bigWig then this: Any Method Of Converting Bigwig File Format Into Bed Format?

Answer 1

[liartom]

Hi! Why don't you use MACS2? As far as I know it generates only bed and bedGraph files in output, i.e. files with start and end coordinates and no WIGs whatsoever

[gracie]

Thanks for answer. I don't have access to raw files. They have only uploaded wig files. Can I still use MACS2?

[Sinji]

Also, you should have access to the raw files if you can find the SRA number associated with the experiment. A quick google search with the SRA number should give you the appropriate GEO database.

[liartom]

Found the problem. You kinda should have access to raw files, as Sinji said, there ought to be GEO accession number and SRA for experiment, so you can redo MACS2 if you need to. The problem with wig is that you almost never can detect which coordinate was used, unless it was explicitly told. I know experiments where people used the middle of genomic interval and I know other cases, where only the first coordinate was used (primarily for microarray data, the start coordinate for tile is often used). If you manage to find SRA you just need to download it and SRA-toolkit, an official ncbi toolkit for processing sra, then you can extract fastq or even bam files, using fastq-dump or sam-dump. Good luck!

[Devon Ryan]

You can if you convert the wig files to bedGraph (they're closely related formats). Note that you might need to tweak the bedGraph values so that MACS2 is doing the right thing with them (the available details are available here, though you'll need to figure out some things yourself I expect).