search

COMBINE-lab / oarfish

long read RNA-seq quantification
BSD 3-Clause "New" or "Revised" License
72 stars 3 forks source link

Incorrect check for aligner #2

Closed dpryan79 closed 9 months ago

dpryan79 commented 1 year ago

I aligned data with minimap2 and made a BAM file out of it with samtools, which unfortunately means that samtools is in the header:

@PG     ID:minimap2     PN:minimap2     VN:2.24-r1122   CL:minimap2 -ax asm5 -t 4 combined.fa reads.fastq
@PG     ID:samtools     PN:samtools     PP:minimap2     VN:1.15.1       CL:samtools view -Sbo alignments.bam -

that results in:

thread 'main' panicked at src/main.rs:254:9:
assertion `left == right` failed: Currently, only minimap2 is supported as an aligner. The bam file listed samtools.
  left: "samtools"
 right: "minimap2"

You might want to check for minimap2, samtools, or pbmm2 (unless it reports minimap2 in the header, I've never checked).

rob-p commented 1 year ago

Hi @dpryan79,

Good catch! Right now, I don't know if we are relying on minimap2 specific flags (I think we are). So I think we still want to make sure minimap2 appears in the list of programs, but perhaps we should say that is must appear among the list, rather than be the first one?

In other words, samtools could be indicative of anything (any alignments later processed with samtools). So we probably don't want to accept just that, but certainly if it's minimap2 post-processed with samtools we should accept it.

--Rob

P.S. Actually, @dpryan79 — I can see that precisely this functionality is implemented on the dev branch ;P. I wonder if you might give this a try and see if it works.