search

jts / nanopore-rna-analysis

14 stars 5 forks source link

Why is p0 required? #2

Open callumparr opened 5 years ago

callumparr commented 5 years ago

Sorry to bother. May you provide a little more background as to why p0 flag, that is too keep all secondary alignments, is required for generating the PAF, and in turn determination the transcript counts?

I read the original Salmon paper but I am not so familiar with computational biology.

In my naive thinking, would secondary alignments make the process of determine abundance more difficult?

Sorry again for stupid question but I am trying to get my head round these things.

jts commented 5 years ago

Hi,

This script implements an EM-based algorithm for transcript abundance like RSEM, Kallisto, salmon, etc. For these methods to work well it is better to keep all alignments of each read, rather than just selecting the "best" one, then let the EM framework resolve the ambiguity. I suggest you look at this paper if you'd like to understand why.

Note this repository is a prototype and not really intended for serious use. We are working on a better implementation that is suitable for use by the community.

Jared