Filtering fastq (by quality score and length), optimum criteria?

Question

[dina.hesham139]

For filtering fastq files (of RNAseq data) (by quality score and length) in galaxy, what are the optimum criteria?

i.e. the min and max size, the min and max quality and Maximum number of bases allowed outside of quality range.

My datasets are from human samples, Hiseq2000, paired end experiment (2 separate files per sample).

Answer 1

[Sean Davis]

Assuming that you are using some alignment process (and not de novo assembly), we generally do not filter and trim tails only very lightly. The alignment process itself is a great filter.

[dina.hesham139]

Thank you alot

Answer 2

[Ian]

As a core facility we generally run our sequences through Trimmomatic to remove adaptor sequence (most important to avoid mapping errors) and trim reads when a moving 4nuc window has a mean quality score below 20.