The bam file used for DiffBind?

Question

[Miao Zhang]

When use DiffBind to compare ATAC-seq data we need bam file for sample sheet. What kind of bam file do we need? Can I use the bam file which already removed duplicates? OR just primary bamfile?

Answer 1

[Rory Stark]

You should NOT remove duplicates, use the full original bam file.

You can chose to remove duplicates when counting reads in DiffBind, however, depending on the type of sequencing you are doing, the duplicates may represent true signal and not only technical duplication. If you are doing relatively long, paired-end sequencing (eg. 150x2) than you may want to set bRemoveDuplicates=TRUE when calling dba.count().

[Miao Zhang]

Thanks, Rory, that helps me a lot!