Extract counts from paired end bam files

Question

[minoo]

I have two bam file for each sample which are paired end sequenced. I need to extract raw counts from these bam files which they look like S1.bam and s2.bam for each sample. I founf these by a search but still don't know how to move forward as they are paired end and two bam files.

featureCounts -T 4 -s 2 \   -a ~/annotations/Homo_sapiens.GRCh38.104.gtf
\   -o ~/results/counts/s1_featurecounts.txt \  
~/results/STAR/bams/*.out.bam

Apparently as they are paired, they need to be sorted by

samtools sort

but I don’t know how to make a single bam file from the two paired that I already have. And how to insert it in this formula. Any help would be appreciated.

Answer 1

[ATpoint]

It's far simpler than you probably think. featureCounts will sort paired-end files as needed if you tell it that the data are paired. You don't need to manually do any sorting or indexing. In fact I never sort files from STAR as featureCounts anyway would sort them back by name in case of paired-end data.

#/ The -p flag tells it to consider the files as paired-end
featureCounts -T 4 -s 2 -a ~/annotations/Homo_sapiens.GRCh38.104.gtf \
  -p \
  -o ~/results/counts/s1_featurecounts.txt \
  ~/results/STAR/bams/*.out.bam

There is not more to do, no merging or whatever.

[minoo]

Thanks for your quick resoponse ATpoint , but how can I add my two paired bam files here in this formula?

[ATpoint]

It's the ~/results/STAR/bams/*.out.bam part, it means "use all bam files". Alternatively replace that with the paths to your bam files.

[GenoMax]

how can I add my two paired bam files here in this formula?

I hope you did not align your paired-end reads independently i.e. Read 1 is in S1.bam and R2 is in S2.bam. That would be incorrect. You have to align paired-end data in a single alignment run.

[minoo]

I didn't align anything. I just put two end paired bam files that I had in a folder and run the above code by changing only .out.bam to .bam. This worked but I don't know if this is correct or not.

[GenoMax]

BAM files are not paired-end but they can have alignments that came from a paired-end dataset. Hopefully that is clear.

[swbarnes2]

Two bams implies two different samples. You could combine them if you are sure they are the same thing, but it doesn't sound like you are sure.