Lima is not orienting sequences to 5’ → 3’ orientation

[ccalix]

According to Iso-Seq Docs, 'Lima will orient sequences to 5’ → 3’ orientation according to the asymmetry of the primers.'. This is definitely not happening with my data and I have no idea why. Any help will be greatly appreciated.
I have the file hifi_reads.bam, where ~23% of reads contain polyt(19) -- therefore probably in the wrong orientation. I check the numbers with:
samtools view hifi_reads.bam | awk '{print $10}' | grep -o 'TTTTTTTTTTTTTTTTTTT' | wc -l
I run lima (2.13.0) with either --isoseq or --hifi-preset ASYMMETRIC

lima --peek-guess --isoseq -j 12 \
  hifi_reads.bam \
  IsoSeq_v2_primers_12.fasta \
  IsoSeq_output.bam

or

lima --peek-guess --hifi-preset ASYMMETRIC -j 12 \
  hifi_reads.bam \
  IsoSeq_v2_primers_12.fasta \
  Preset_output.bam

Both lima runs are able to trim the primers (IsoSeqX_bc03_5p and IsoSeqX_3p) really well. I also manually inspect this with:
samtools view IsoSeq_output.bam | head -n 10
Even the reads with polyt at the 5'end have the primers removed. Before trimming, these polyt-containing reads had the reverse complement of these primers, confirming their wrong orientation. But orientation is not changed after lima! There are still ~23% of reads with polyt(19) in both IsoSeq_output.bam and Preset_output.bam.

Can anyone share with me their results for the amount of polyt reads after lima? Why is this not working for me?

[jandubois]

You ended up in the wrong GitHub repo. To file a bug against the bioconda software try the https://github.com/PacificBiosciences/pbbioconda issue tracker.