Using different quality thresholds for trimming and filtering reads

Question

[samybamy_]

I'm performing de novo assembly from my transcriptome data, and using the FastX toolkit to trim and filter my reads.

There is a -t parameter for trimming and -q parameter for filtering, both of which are quality thresholds. Are they conceptually different? More specifically, would it be silly to set t=30 while trimming and then q=22 while filtering, since t=30 is a more stringent cut off, making q=22 redundant? Does it make sense to only set the -t and -q parameters to be the same cutoff, or is there a benefit in setting t=22 while trimming, and then making it more stringent while filtering, with q=30?

I'm super new to this.

Thanks! Sam

[lh3]

No, don't apply quality trimming or filtering for de novo assembly. Doing that often reduces the quality of assembly. Mainstream assemblers come with their own preferred approaches to deal with low-quality reads. Also, don't use fastx toolkit nowadays. It is outdated by recent trimming and filtering tools which are much faster and use better algorithms.