featureCounts for paired-end reads

Question

[xosugarxspicexo]

Below is the samtools flagstat output for my BAM file, which only contains mapped reads (as seen from 100.00% mapped)

I know featureCounts by default does not consider/assign multimappers (i.e., secondary). Does it also skip over singletons/unproperly paired reads? Is it correct to say they only consider primary, uniquely mapped, properly-paired reads for assignment?

In that case, is there any way to tell from the below output how many reads that is? There seems to be 9539764 / 2 = 4769882 properly paired reads, but I'm assuming some of these are multimappers? So it would be less than that number of reads that featureCounts attempts to assign a genomic feature?

20155191 + 0 in total (QC-passed reads + QC-failed reads)
9566830 + 0 primary
10588361 + 0 secondary
0 + 0 supplementary
0 + 0 duplicates
0 + 0 primary duplicates
20155191 + 0 mapped (100.00% : N/A)
9566830 + 0 primary mapped (100.00% : N/A)
9566830 + 0 paired in sequencing
4788364 + 0 read1
4778466 + 0 read2
9539764 + 0 properly paired (99.72% : N/A)
9539764 + 0 with itself and mate mapped
27066 + 0 singletons (0.28% : N/A)
0 + 0 with mate mapped to a different chr
0 + 0 with mate mapped to a different chr (mapQ>=5)

[GenoMax]

Assuming you are using new version of featureCounts did you specify -p --countReadPairs in your command line?

To count fragments (instead of reads) for paired-end reads, the --countReadPairs parameter should also be specified.

[xosugarxspicexo]

I'm using v.1.5.2, which I think doesn't have the new options, so my command looks something like:

featureCounts -pC -g gene_id -t exon -a annotation.gtf -s 0 -o counts.txt sample1.bam sample2.bam

[Istvan Albert]

note how featurecounts operates and counts reads over various intervals, that typically cover a subset of the genome. Hence you would expect to match fewer reads than what is in a BAM file.