classify bins catches on weird taxa

[karoraw1]

Apparently there is a bit of hamster in my mapping.tax file.

An unrecoverable error occurred: std::exception

Here is some debugging information to locate the problem:
/home/johdro/projects/taxator-tk_default.git/src/fileparser.hh(52): Throw in function FileParser<FactoryType>::RecordType* FileParser<FactoryType>::next() [with FactoryType = AlignmentRecordFactory<AlignmentRecordTaxonomy>; FileParser<FactoryType>::RecordType = AlignmentRecordTaxonomy]
Dynamic exception type: boost::exception_detail::clone_impl<Exception>
std::exception::what: std::exception
[exception_tag_line*] = 1488404
[exception_tag_taxid*] = 10026
[exception_tag_general*] = bad alignment reference taxon

I fixed this by using

taxknife -f 2 --mode traverse -r species genus family order class phylum superkingdom < mapping.tax > newmapping.tax

as referenced here:
fungs/taxator-tk#51

i am using an older version of metawrap so i am not sure if you've corrected for this already, but i thought i would pass it along.

[ursky]

Thanks for pointing it out. I am a bit confused about your fix though - metawrap runs the command cat ${out}/binned_predictions.txt | taxknife -f 2 --mode annotate -s path | grep -v "Could not" | cut -f1,2 > ${out}/contig_taxonomy.tab. This returns the taxonomy of each contig. How would you use your solution to fix it?

Looking at the error, I wonder if its caused earlier versions of metawrap not installing the correddct gcc/boost libraries (I had quite a few error related to that). If you dont mind, can you install metawrap in a new conda environment (when it comes to library dependencies updating doenst always fix it) and see if it still gives you the error?

[karoraw1]

sorry for not explaining the issue well enough. the error is produced earlier in the script by the taxator command (line 147).

the megablast results yielded a mapping.tax file that contained the taxid 10026. according to the nodes.dmp file in the NCBI taxonomy this is a "subfamily". taxator only accepts the major taxonomic hierarchy levels listed in the taxknife command i showed above. I used taxknife to remove it and then re-ran taxator on the new mapping file, without issue.

i've been meaning to update anyways, so I will go ahead and get that started.

[ursky]

Ah, I understand now. I encountered similar issues before. Actually, I already have a dedicated script bin/metawrap-scripts/prune_blast_hits.py to sort out the "no rank", "subspecies", "varietas", "forma" from the blast hits before proceeding. I never encountered a "subfamily" before, but I guess Ill just add it to the exception list. This fix should be in metaWRAP v=0.9.3 when it rolls out!