Alignment and Counting strandedness redundancy?

Question

[brismiller]

Hi everyone I have a question about the parameters to use for my SE RNA-Seq samples for the alignment and counting steps of my pipeline.

I am using hisat2 and featureCounts for the alignment and counting programs, upstream of DESEQ2 for detecting differential expression. I know that my reads are reversely stranded, not only from the protocol for library preparation but also because if I run hisat2 with the [--rna-strandedness "F"] option the bam file viewed via igv shows almost all of the reads pointing in the opposite direction as what the annotation file gtf shows.

My question is this: If I run hisat2 [--rna-strandedness "F"] and then featureCounts [-s 2], I get the same results as if I do hisat2 [--rna-strandedness "F"] and then featureCounts [-s 2]. But wouldn't the "R" and -s 2 options counteract each other when used in the same pipeline and thus the reverse complement of the read would not be used for counting?

Answer 1

[h.mon]

The --rna-strandness just set the XS tag:

With this option being used, every read alignment will have an XS attribute tag: '+' means a read belongs to a transcript on '+' strand of genome. '-' means a read belongs to a transcript on '-' strand of genome.

However, featureCounts does not use this tag, it relies on sam flags to guess strandedness.

[brismiller]

Thank you h.mom this answers my question!

For my own sanity, I made sure that the two sample aligned with "F" or "R" had the same number of reads with the 16 flag (for read being on the reverse strand). They were the same as anticipated.

#sam made using --rna-strandness "R"
samtools view -c -F 16 Test_sample_Revere_alignment.bam 
2793953

and then

#sam made using --rna-strandness "F"
samtools view -c -F 16 Test_sample_Forward_alignment.bam 
2793953