HT-seq - what is ideal?

Question

[jsl]

Hi guys,

I ran a HT-seq command and I would like to cross check with you guys. wonder what output is ideal? Below is the command

htseq-count -f bam --idattr=gene -r pos /home/user/scratch60/STARresults/SRR7059136Aligned.sortedByCoord.out.bam /home/user/scratch60/NCBI_files/GCF_000001405.26_GRCh38_genomic.gff >/home/user/scratch60/HTseq_annotation/annotated_SRR7059136.txt

and the output is this

12600000 SAM alignment record pairs processed.
Warning: Mate pairing was ambiguous for 22805 records; mate key for first such record: ('SRR7059136.1152992', 'first', 'NC_000001.11', 135867, 'NC_000001.11', 493007, 357290).
12621898 SAM alignment pairs processed.

My questions are:

1. Should I be concerned about missing mate encountered warnings? Is there an ideal number one should be aiming for?
2. Am I right to run -r pos because my STAR command included --outSAMtype BAM SortedByCoordinate? I'm trying to understand the logic of this, if someone can explain it, it would be much appreciated!

Thanks guys!

Answer 1

[Devon Ryan]

1. In an ideal world it'd be 0, but losing <1% won't affect anything.
2. Sure, though you can just have STAR quantify things for your and not have to wait as long.

In general HTSeq-count isn't much used these days because it's quite slow. Either have STAR do the counting for you or use featureCounts and you'll get the results quicker.

[jsl]

Thanks Devon! That makes sense.