potential output problem

[retrogenomics]

Hi,
Using cutadapt like this (1.9dev1):
cutadapt -e 0.12 -m 25 --info-file="info.tab" --too-short-output="discarded_tooshort.fastq" --untrimmed-output="discarded_missing_adapter.fastq" -g $linker "trimmed_bc.fastq" -o "trimmed_linker.fastq" > "info.log"

gives me this output:
=== Summary ===

Total reads processed: 5,022
Reads with adapters: 4,999 (99.5%)
Reads that were too short: 392 (7.8%)
Reads written (passing filters): 4,630 (92.2%)

Total basepairs processed: 908,639 bp
Total written (filtered): 693,472 bp (76.3%)

However, when I count the number of reads in "trimmed_linker.fastq", I actually get 4,607 (4,999-392). What are these extra 23 reads counted in the 'Reads written (passing filters)' ?

[marcelm]

The 23 reads are the (5022-4999) untrimmed ones. Since they are also written to an output file (--untrimmed-output), they are also counted under the "Reads written (passing filters)" heading.

I agree this is confusing and should be changed, in particular because the reads in --too-short-output are not counted here. Perhaps my recent simplification of the statistics summary simplified things a bit too much. I’ll try to change this soon. Until then, you can calculate "Reads written (passing filters)" minus ("Total reads processed" minus "Reads with adapters") to get the number of reads written to the output file.

[retrogenomics]

Actually, I still get the same 4630 number of reads, whether I use the --untrimmed-output (or the --too-short-output) option, or not.

[marcelm]

If you don’t use --untrimmed-output, then both trimmed and untrimmed reads are written to the regular output file, so the number does not change (you should now have 4630 reads in trimmed_linker.fastq). The number should change when you use --discard-untrimmed.

[retrogenomics]

yes indeed ! Thank you for these explanations.

[marcelm]

I’ve opened issue #128 to track this. Thanks for reporting.