Does trinity discard reads shorter than 25-mer?

Question

[lwc628]

I was curious if Trinity discard reads shorter than its default 25mer.

The reason why I asked is after quality checking and trimming, there are reads in my fastq files that are shorter than 25. I wanted to know what would happen to them in Trinity.

I assume they(reads shorter than 25 bases) would get discarded as they would not be included in the 25-mer library, but wanted to confirm this

Answer 1

[Damian Kao]

It would not be used in the Inchworm step as it uses jellyfish to hash the 25-mers. Jellyfish would just skip reads < 25. However maybe it is used in the mapping back steps in chrysalis. I would ask this on the official Trinity mailing list. The developers are usually pretty fast at replying to questions.

Answer 2

[lwc628]

I posted this question in the Trinity user group, and here is what Mr. Haas responded to my question:

I'm pretty sure that reads shorter than 25 will simply be ignored. I know they'll be ignored by the kmer counting step (initial part of Trinity), and pretty sure they'll be ignored later on by the Chrysalis/Butterfly stage.

Note, unless your reads are at least ~50 bases in length, for the majority of your reads, Trinity might not be the best application for you. You can still give it a whirl, though. Definitely try others too, for sake of comparison.

best,

~brian