Funannotate fix 'too many values to unpack"

[PlantDr430]

I finished a prediction on an older version (1.3.2) and got a message about needing to fix a part of the .tbl output file. I've done this before and have used the "fix" command successfully before, but now when I am trying to run it I get an error of "ValueError: too many values to unpack".

stephenwyka@bspmgenomics:/data/wyka/funannotate/Batch_2/Cfusi$ funannotate fix -i Cfusi_fun_output/predict_results/Claviceps_fusiformis_PRL_1980.gbk -t Cfusi_fun_output/predict_results/Claviceps_fusiformis_PRL_1980.tbl

[10:28 PM]: OS: linux2, 24 cores, ~ 264 GB RAM. Python: 2.7.12
[10:29 PM]: Running funannotate v1.3.2-3ba4c2e
Traceback (most recent call last):
File "/data/linuxbrew/.linuxbrew/Homebrew/Cellar/funannotate/1.3.2/libexec/util/updateGBK.py", line 62, in
locustag, genenum, justify = lib.getGBKLocusTag(args.input)
File "/data/linuxbrew/.linuxbrew/Homebrew/Cellar/funannotate/1.3.2/libexec/lib/library.py", line 2293, in getGBKLocusTag
tag, count = lastTag.split('_')
ValueError: too many values to unpack

[nextgenusfs]

What does your locus_tag look like? It is expecting a single underscore. So it seems that it found multiple underscores in the name.

[PlantDr430]

Oh okay, I manually changed it as I wanted two underscores in the locus tag. I guess I'll have to change it back. If you get the time could you also look at this SNAP gff3 file that I have. It took a while to figure out how to get SNAP outputs into gff3 format. I finally did, but now when I use the --other_gff (filename):(weight) it tells me that the gff3 file is not a valid file. I've attached the file as a .txt

stephenwyka@bspmgenomics:/data/wyka/funannotate/snap_runs/LM4_snap$ nice -n 19 /data/bin/funannotate-1.4.0/funannotate predict -i LM4A_assembly_masked.fasta -o LM4C_longaugustus_fun_output -s "Claviceps purpurea" --isolate LM4C --protein_evidence ../../../../../home/stephenwyka/Pconf.fasta --cpus 20 --busco_seed_species Claviceps_purpurea --busco_db sordariomyceta_odb9 --name LM4C\_FUN --min_protlen 100 --optimize_augustus --soft_mask 1000 --other_gff LM4_snap.gff3:1 
-----------------------------------------------------------------
[09:48 AM]: OS: linux2, 24 cores, ~ 264 GB RAM. Python: 2.7.12
[09:48 AM]: Running funannotate v1.4.0-90e1680
[09:48 AM]: AUGUSTUS (3.2.2) detected, version seems to be compatible with BRAKER and BUSCO
[09:48 AM]: LM4_snap.gff3:1 is not a valid file, exiting

LM4_snap.gff3.txt

[nextgenusfs]

Sure I will take a look. There should be some snap2gff scripts out there @hyphaltip do you recall any? I will make this script support the multiple underscores, it will just split on the last one. Probably get it updated tomorrow.

[PlantDr430]

yea I actually had to do a zff to gff3 (which wasn't gff3) , but then found a converter from that gff to an actual gff3 file (which is the EvidenceModeler's, SNAP-to-gff3.pl).

[nextgenusfs]

The --other_gff files need to pass EvidenceModeler validator, that script is located in $EVM_HOME/EvmUtils/gf3_gene_prediction_file_validator.pl. So run that script on your GFF3 file and see what errors it is giving you. I'd be surprised that the SNAP-to-gff3.pl isn't give you compatible output?

[PlantDr430]

That's what I thought. Especially since the script comes with EvidenceModeler. I'll run the validator later and see where that gets me. Thanks for the help.

[PlantDr430]

So I ran the validator and got no errors.

swyka430@swyka-ubuntu:/Wyka/bioinformatics/LM4$ perl ../../../usr/local/bin/EVidenceModeler-1.1.1/EvmUtils/gff3_gene_prediction_file_validator.pl LM4_snap.gff3 
swyka430@swyka-ubuntu:/Wyka/bioinformatics/LM4$ 

I also ran it with an incorrect gff version to make sure it was working and I got errors.

swyka430@swyka-ubuntu:/Wyka/bioinformatics/LM4$ perl ../../../usr/local/bin/EVidenceModeler-1.1.1/EvmUtils/gff3_gene_prediction_file_validator.pl LM4_snap_wronggff3.gff3 
Use of uninitialized value $feature_info in pattern match (m//) at ../../../usr/local/bin/EVidenceModeler-1.1.1/EvmUtils/gff3_gene_prediction_file_validator.pl line 54, <$fh> line 3.
Fatal Error: cannot parse ID from entry
contig_1	snap	CDS	7191	7196	.	-	Name=contig_1-snap.1 at ../../../usr/local/bin/EVidenceModeler-1.1.1/EvmUtils/gff3_gene_prediction_file_validator.pl line 54, <$fh> line 3.

So I don't know why Funannotate is having trouble with the correct LM4_snap.gff3 file.

I'm going to try another validator (GenomeTools) and see if that helps.

[nextgenusfs]

So it seems to be passing for me locally as well, what was the command that you issued to funannotate predict and then the error that it gives you? Maybe I'm looking in the wrong place.

[nextgenusfs]

Oh... I think I see the problem. Did you pass a weight for this file, i.e. --other_gff LM.gff3:1?

[PlantDr430]

Yea, I ran a weight --other_gff LM4_snap.gff3:1 you can see it at the end of my command line in the previous post.

[nextgenusfs]

Okay, da6734e should fix it. The error was that the file couldn't be found in a previous check because there is a colon in the name (that needed to be fixed).

[PlantDr430]

Okay, thanks for looking into this. I'll let me admin know to update to the latest version once a new update comes out.

[nextgenusfs]

In the meantime it should work to just leave off the :1, it will default to a value of 1 automatically.

[PlantDr430]

Cool, I'll give that a shot