low Successfully assigned alignments

Question

[Srujana]

Hello,

I am doing RNAseq analysis, i got above 90% alignment score in mapping step using HISAT2, but when i am trying to do quantification using featureCounts , i am getting low successfully assigned alignment scores around 50%. What could be the reason for that? My data is paired end and i am using the following command

i have also tried using unsorted bam file but same result

featureCounts -p -M -s 2 -O -T 8 -a /path/to/refgen/Mus_musculus.GRCm39.112.gtf -o /path/to/output/featureCounts_output.txt  /path/to/sorted_bam_files/*.sorted.bam*

Any help in this will really appreciated.

[GenoMax]

Check to see if you have rRNA annotations in your GTF. It is possible that your data still contains rRNA and was not counted.

[Srujana]

Thankyou for the reply. I have added -t gene in the command ,now i am getting 85% score.

Answer 1

[Carlo Yague]

-s 2 assumes a RF orientation for your library and might not be appropriate depending on your library type and mapping. So you can try replacing -s 2 by -s 0 (unstranded) or -s 1 (FR orientation) in case you got the strandness wrong.

If this doesn't work, I recommend loading the indexed alignment file (.bam and ..bai) directly into a genome browser such as IGV and verify whether (1) the read pairs behave as expected (orientation, distance, etc...) and (2) if the reads fall onto annotated region but are depleted from introns or intergenic regions.

[Srujana]

Thankyou for responding. -s 2 and -s 0 are giving the same score. i have added -t gene in command, now i am getting around 85% score, i hope now its correct.

[Carlo Yague]

Hi, thank you for this additional information. By default, featureCounts will consider only the exons from the provided gtf file. If using -t gene increases the mapping rate from 50% to 85%, that means that 35% of your reads map on portions of genes that are NOT exons (so probably introns ?)

IMHO 35% of reads on introns is a little bit worrying. But perhaps it is just that the annotation (.gtf) is not complete and misses exon information for some genes.