featurecounts vs samtools view

Question

[tonja.r]

I needed to check the results coming from featureCounts from subread. In fact, samtools outputted less reads then featurecounts did.

I run samtools as follows:

samtools view -c -q 1 -F 4 -L matrix_gene.bed file.bam
2732259

And featureCounts:

featureCounts --minReadOverlap 50 -s 1 -Q 1 -T 12 -a matrix_gene.saf -F SAF -O -o $OUT/total_counts.txt file.bam
featureCounts --minReadOverlap 50 -s 2 -Q 1 -T 12 -a matrix_gene.saf -F SAF -O -o $OUT/total_counts.txt file.bam

​

(minReadOverlap = 50 because the length of my reads is 50, -s 1 for stranded and -s 2 for reversely stranded, -Q mapping quality and -O for multiple overlap (as samtools outputs alignment multiple times if they overlap multiple regions)

samtools results: 2732259

featureCounts -s 1: 1661976
featureCounts -s 2: 1662917

So, the number of stranded and reverse stranded reads: 3324893

Why do the results differ if I specified the same parameters in both methods?

PS: I did not find what overlap between a read and a region samtools allows, but if the overlap less than the read length, samtools should have more reads than featureCounts.

Answer 1

[Istvan Albert]

The answer to this type of questions usually very simple - the two tools actually do different things. So the discrepancy is unlikely to be caused by software errors.

Of course why do they do different things - now that is harder to answer. Make sure to investigate the preconceptions and tacit assumptions that you have about the parameters and what the tools actually they do.

For example a read may overlap with features on both positive and negative strands - in which case it will be double counted. Etc.