search

BrendelGroup / AEGeAn

Integrated toolkit for analysis and evaluation of annotated genomes
http://brendelgroup.github.io/AEGeAn
ISC License
24 stars 10 forks source link

Problem running ParsEval: Cannot infer missing exons for mRNA #117

Closed damianosmel closed 10 years ago

damianosmel commented 10 years ago

I run ParsEval, resulting to the following error:

"cannot infer missing exons for mRNA 'badA1.t01' (line 579) without CDS feature(s)
Assertion failed: (agn_typecheck_gene(fn)), function gene_stream_next, file src/core/"

I am running Parseval through AEGeAn v. 0.11

Thanks!

standage commented 10 years ago

Could you send me the data files you used while running ParsEval? Or at least a small portion of each file that will reproduce this error? It's hard to tell whether this is a problem with the program or with the input files without actually seeing the files.

Also, this tidbit from the ParsEval documentation is relevant.

damianosmel commented 10 years ago

Sure, how I can find your email?

standage commented 10 years ago

It's included a few places in the documentation, but I'll include it here for convenience: daniel.standage@gmail.com.

damianosmel commented 10 years ago

Thanks, I send you an email.

standage commented 10 years ago

I cannot reproduce this error. When I run ParsEval on the entire file, it gives me the previous error you reported (ID clash). When I isolate the feature listed above, it gives me no error. You will need to provide more details if you would like me to help identify and solve the issue.

damianosmel commented 10 years ago

I see. After resolving the ID clash issue by assigning the gene "gltX" to only one strand (e.g both instances to "+" strand), then running ParsEval I encounter the current issue. Is it possible to try this modification in the gff and run again the tool? (or I can send you the modified gff). Thank you for your time.

standage commented 10 years ago

I think you had better send me your modified GFF3 file. My impression is that you have modified it in other ways as well. If you fix or remove the gltX problem, the same exact problem shows up with a different ID. I still cannot reproduce the "cannot infer missing exons" error message.

damianosmel commented 10 years ago

I sent you the modified gff. Thanks.

standage commented 10 years ago

Okay, it looks like this file is absolutely full of problems. In the latest file you sent, you manually changed the strand of two features, but this doesn't really fix the real problem. The real problem is that in many cases you have different features using the same ID. The gltX ID is just one example, but the same thing is happening all over the GFF3 file. Two distinct features (2 different genes for example) cannot have the same ID, or it is going to cause all kinds of problems. I'm guessing that once you fix this problem with the data the error message will go away, so I'm going to close this thread. Feel free to reopen it if the error persists after the data file is fixed.