Question: sorted bam files nanopore data

[rebeelouise]

Just out of curiosity, with single read data that has a 3' bias on the RNA strand, what would be the impact of sorting bam files with samtools before quantification with salmon?

I have used salmon on sorted and unsorted bams (out of curiosity) and the data when imported into R looks "better" when it comes from sorted bam files.

By better I mean, more significant transcripts identified that are changing in abundance and my experimental groups cluster well.

Obviously, you can't do something just because it looks "better", but I wondered if you had any insights on to how this would impact quantification.

The data has been generated by cDNA-PCR sequencing on nanopore. I don't always get full length transcripts due to sample type and due to the method you only acquire the 3' end of the transcript.

Thanks in advance and look forward to discussion/answers!!