CDS-extractor.pl should return error message (and exit code 1) when no CDS could be extracted

[jvollme]

When genbanks are used as input that still have windows/dos line-endings, cds-extractor.pl just quits without an error message, giving the impression that it successfully extracted all CDS.

Maybe it could either be adjusted to tolerate windows line endings or to always double-check the number of extracted CDS when finished and generally raise a warning if it is zero.

[jvollme]

Ah, nevermind. The ACTUAL problem was that the genbank had ALL cds marked as "/pseudo". Had not expected that :)

But here's another thought: Maybe you could include an option to output pseudogenes also?
A general warning (and optimally exit-code 1) in cases were zero CDS were extracted would still be very helpful though, for including the script in automized pipelines.

[aleimba]

Hi @jvollme,

thanks for bringing up the issue!

1. Windows <-> Unix line ending issues are very annoying, that's why I recommend running dos2unix on the files https://github.com/aleimba/bac-genomics-scripts#windows---unix-linebreak-problems. And of course it would be best to work only in Unix 😉
2. cds_extractor is designed to ignore /pseudo CDS, because it has several downstream issues. I rather trust more sophisticated ORF-finding software than translate everything by myself, including stop codons etc.. Also for protein extraction I use the tag /translation which is not included for pseudo-CDS. If you want to extract also pseudo genes I recommend using other software, e.g. FeatureExtract from the CBS (http://www.cbs.dtu.dk/services/FeatureExtract/download.php) or extractfeat from the EMBOSS package (http://emboss.sourceforge.net/apps/release/6.6/emboss/apps/extractfeat.html).
3. I have a check in cds_extractorif there is no annotation in the input file at all and quit with error. A file where all CDS have a /pseudo tag is a rather unusual case, but its a good idea to include a "zero CDS extract" error. Will include it.