This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Merge bigwig Files with different intervals

Hello everyone, I am trying to merge about [700 ATAC seq][1] bigwig files together and calculate mean read counts in all intervals. these bigwig files are already normalized , however intervals are not the same in every file. I was wondering if I could calculate read counts in intervals ( coverages) using UCSC bigWigMerge .
Thank you

bigwig merge

Yes, I think so. Have you tried it ?

does bigWigMerge simply add read counts in intervals or does it report the mean read count ?

By default, it reports the sum, but there are other options.

bigWigMerge v2 - Merge together multiple bigWigs into a single output bedGraph.
You'll have to run bedGraphToBigWig to make the output bigWig.
The signal values are just added together to merge them
usage:
   bigWigMerge in1.bw in2.bw .. inN.bw out.bedGraph
options:
   -threshold=0.N - don't output values at or below this threshold. Default is 0.0
   -adjust=0.N - add adjustment to each value
   -clip=NNN.N - values higher than this are clipped to this value
   -inList - input file are lists of file names of bigWigs
   -max - merged value is maximum from input files rather than sum

Deeptools bigwigcompare allows many different operations (log2, ratio, subtract, add, mean, reciprocal_ratio, first, second), along with the optional use of pseudocounts and/or scaling. This is my favorite tool to play with bigwigs.

But that only goes for two files, does it?

pt.taklifi Do you really have 700 bigwigs? What kind of experiment is that, it sounds excessive. Not sure how tools that exist will handle such loads of data.

Hi , ATpoint! Im sorry i was not clear about this. There are 796 ATACseq samples from different cancers and I want to merge all samples from a specific cancer with each other, for example I will merge 141 breast cancer bigwigs with each other. after that I will use mean signal intensity for my project

0 answers

No answers yet.

Log in to answer this question.