From coverageBed to a list of reads per genes

Question

[Bnf83]

Hi guys, I'm analysing acChipSeq data. I performed all the steps of a canonical best practice pipeline. The last step was the coverageBed. Now, since I would like to perform a differential analysis for example using edgeR I need to have a matrix in which rows are genes and columns are samples. The output of the coverageBed is a file that looks like this:

 chr1   9884    10032   0   0   148 0.0000000
 chr1   191374  191638  0   0   264 0.0000000
 chr1   270604  271602  0   0   998 0.0000000
 chr1   272675  272855  0   0   180 0.0000000

It is a .bed file. Since I cannot use FeatureCount with bed files I converted to .bam but they are always corrupted. Any suggestion to have a classical table with genes as rows and samples as columns as FeatureCount does with canonical RNA seq data analysis?

Thank you in advance

Answer 1

[rrbutleriii]

The reason the bam file is not working is because the bed file doesn't contain enough information to re-create reads. Perhaps there is a tool that attempts to reconstruct reads based on coverage beds, but that sounds...not good. You need a true bam/sam for featureCounts.

I am unfamiliar with the acChipSeq protocol, but standard ChipSeq uses a mapping tool to generate a bam/sam at some point upstream of the bed file generation. You want that intermediate file.