Camparee will silently fail to compile the gene_quant file when gtf contains special characters

[geoleven]

Hello,

I came across this problem where:
I have an annotation file created by your own util (from my gtf) which has colons (:) in the names of the gene_id row.
Camparee will run until the last step at which it fails because it opens the gene_quant and another file which only containing one line of entries with no gene_id (but a pre-occuring tab as if an empty gene_id was used).
Following the gtf standard the use of colons is not strictly prohibited so i report this here.
I have positively verified that removing colons and other special characters from the gene_id and transcript_id row fixes the problem.

Suggestions:
Proper fix: quantification step is minimally adjusted to properly handle the tab delimited file (as per your own spec).
Hack fix: if the user goes with the official util to convert the gtf, the non passing special characters are converted to something that is not causing problems (_ maybe?).

I would go with the proper fix as it is both "proper" and probably/seemingly easier to implement as you don't have to take every different possibility in mind. It will also work for users that made their own files eliminating issues like this being reported (as the use of the util is neither mandatory nor suggested in the basic readme).

Feel free to ask me if you need more info.

Best regards,
Georgios Levenits.

[brainfood]

Hi Georgios,
Thanks for your interest in CAMPAREE and for your detailed feedback. Would it be possible for you to include a few lines from the GTF that was originally causing the problem? That will help me get a little more context for the bug. We're in the process of putting together a large update for CAMPAREE, so your suggestions for a fix are both extremely useful and well-timed. Thanks again.

[geoleven]

Sorry for the late response.

Here an example of some lines of the relevant txt annotation file :

1 + 430418 431458 1 430418 431458 L1PA15-16.1.chr1:430417-431458.+ L1PA15-16.1.chr1:430417-431458.+ None None
1 + 440936 447357 1 440936 447357 L1PA7.1.chr1:440935-447357.+ L1PA7.1.chr1:440935-447357.+ None None
1 + 666159 667201 1 666159 667201 L1PA15-16.1.chr1:666158-667201.+ L1PA15-16.1.chr1:666158-667201.+ None None
1 + 675912 682333 1 675912 682333 L1PA7.1.chr1:675911-682333.+ L1PA7.1.chr1:675911-682333.+ None None
1 - 885341 887183 1 885341 887183 L1PA7.1.chr1:885340-887183.- L1PA7.1.chr1:885340-887183.- None None
1 - 4864392 4867230 1 4864392 4867230 L1PA4.1.chr1:4864391-4867230.- L1PA4.1.chr1:4864391-4867230.- None None

This was the result from a gtf in this form:
chr1 Unknown transcript 33129 34441 0 - . gene_id "L1PA6.0.chr1:33528-34041.-"; transcript_id "L1PA6.0.chr1:33528-34041.-";
chr1 Unknown transcript 48017 49918 0 + . gene_id "L1PA14.0.chr1:48416-49518.+"; transcript_id "L1PA14.0.chr1:48416-49518.+";
chr1 Unknown transcript 64026 65066 0 + . gene_id "L1PA4.0.chr1:64425-64666.+"; transcript_id "L1PA4.0.chr1:64425-64666.+";
chr1 Unknown transcript 71785 72925 0 - . gene_id "L1PA7.0.chr1:72184-72525.-"; transcript_id "L1PA7.0.chr1:72184-72525.-";
chr1 Unknown transcript 72126 73297 0 + . gene_id "L1PA7.0.chr1:72525-72897.+"; transcript_id "L1PA7.0.chr1:72525-72897.+";
chr1 Unknown transcript 72609 74078 0 - . gene_id "L1PA16.0.chr1:73008-73678.-"; transcript_id "L1PA16.0.chr1:73008-73678.-";
chr1 Unknown transcript 73445 75295 0 - . gene_id "L1PA2.0.chr1:73844-74895.-"; transcript_id "L1PA2.0.chr1:73844-74895.-";
chr1 Unknown transcript 74498 77097 0 + . gene_id "L1PA2.0.chr1:74897-76697.+"; transcript_id "L1PA2.0.chr1:74897-76697.+";

All these are derived from the annotation file found here:
https://github.com/FenyoLab/L1EM/blob/master/annotation/L1EM.400.bed
(We are trying to modernize their pipeline from using flux to camparee+beers2.)

Should you need anything else ask me, I will try to be faster this time.

Btw totally unrelated but in the code there is somewhere a comparison of string using "is" which may cause troubles. If I find the time I will do a minor pull request on that one.

[brainfood]

Hi Georgios,
My turn to apologize for a delayed response. I'm locked down with other work at the moment, but I'm dropping this comment here to acknowledge I see your response and will start working on it once I've freed up some time. Thanks again for your detailed feedback!

[brainfood]

Hi Georgios,
Very sorry for the long delay on this issue. I was finally able to carve out some time for the fix. If this is something you're still interested in, you can test the fix in commit b839bbb . If you do test it, please let me know how it goes. Thanks again for raising this issue and apologies for the long silence on my part.