Clarification on quality filtering/trimming

[enricoferrero]

Can you please clarify here and in the documentation or help files at what stage the quality filtering/trimming is done? Is it before or after the adapter trimming?

From the help file:

 Filtering & Post-trimming & Redistribution:
          -q, --end-quality  <int> Trim 3' end until specified or higher quality reached; (0)
          -Q, --mean-quality <int> The lowest mean quality value allowed before trimming; (0)

From the output:

8000000 read pairs processed; of these:
     23 ( 0.00%) read pairs filtered out by quality control
   4199 ( 0.05%) short read pairs filtered out after trimming by size control
    175 ( 0.00%) empty read pairs filtered out after trimming by size control
7995603 (99.95%) read pairs available; of these:
4834307 (60.46%) trimmed read pairs available after processing
3161296 (39.54%) untrimmed read pairs available after processing

Thanks!

[relipmoc]

The mean quality filtering is before adapter trimming while the end quality trimming if after adapter trimming. Thank you for your advice! I'll add this clarification to the skewer manual.

[nick-youngblut]

Sorry to hit up an old post, but it could help to have some clarification on the documentation:

 Filtering & Post-trimming & Redistribution:
          -q, --end-quality  <int> Trim 3' end until specified or higher quality reached; (0)
          -Q, --mean-quality <int> The lowest mean quality value allowed before trimming; (0)

For "The lowest mean quality value allowed before trimming", do you mean "...allowed before filtering"? It's not clear, at least to me, how --end-quality is different from --mean-quality, base on how the documentation is worded. It seems like --mean-quality means: "if the mean quality of the read is < the cutoff, then filter the read", but the wording appears to be otherwise ("trimming"). Is there a way to filter a read (no trim) based on average read quality?