Is HISAT2 able to output only the best alignment?

Question

[RoseString]

I am studying a repetitive region in which multiple lncRNAs are transcribed. For now, I am following the HISAT2-StringTie pipeline to assemble and quantify novel lncRNAs from this region. Because of the repetitive nature of this region, I only want to keep reads that are

1. not strictly uniquely mapped because there would be very few aligned reads left; and
2. assigned to their best possible locations.

This is the default behavior for aligners like bowtie2 (in the manual, "Default mode: search for multiple alignments, report the best one"), but I did not find the options to do so in HISAT2 (in the manual, "Default mode: search for one or more alignments, report each"). The closet thing that I can find is to set -k to 1, but that does not guarantee that the alignment is the BEST. I also cannot filter alignments based on their MAPQs, because there are only three MAPQ values available:

• 0 (unmapped)
• 1 (not uniquely mapped); and
• 60 (uniquely mapped).

Does anyone has experience to identify the best alignment using HISAT2? Thanks!

Answer 1

[prasundutta87]

When you say 'best alignment', I believe you mean 'primary alignment'. According to the HISAT2 manual (https://ccb.jhu.edu/software/hisat2/manual.shtml), 'Primary alignments mean alignments whose alignment score is equal or higher than any other alignments. It is possible that multiple distinct alignments have the same score'. In the case of equal scores, the primary alignment is assigned arbitrarily. You can remove non-primary alignments using parameters –F 256 in samtools view. Using the flag parameter seems to be the only option to get the 'best alignment'.