Using built HMMs or from COGs about metatag

[Robaina]

Alright, closing the issue then!

[Robaina]

Hi there,

According to hmmsearch's error messages, it seems like you could change hmmsearch's parameter --cut_nc by --cut_ga (or drop it altogether if that doesn't work). These parameters control the model-specific score thresholding, according to hmmsearch documentation (pag. 76).

To modify the setting, go to makedatabase.py and change line 86, additional_args='--cut_nc' by additional_args='--cut_ga' or additional_args=None:

https://github.com/Robaina/TRAITS/blob/ea8aaf3484c6f3e70ae384adee43cbf62af48872/code/makedatabase.py#L81-L87

The previous are additional arguments to hmmsearch, so any valid argument to hmmsearch can be added this way.

If that doesn't work we will consider other options :)

[gecko1990]

It worked well, although the problem was that you have to write the value twice:

TC 207.409 207.409

I added at the end the value from the COG webpage as the TC value. According to the HMM documentation (page 76.), the "TC thresholds are generally considered to be the score of the lowest-scoring known true positive that is above all known false positives." This seems to "correspond" to the definition of the value COG webpage, the Threshold Bit Score defined as:

"the domain specific threshold score (shown as a bit score) that an RPS-BLAST hit must meet or exceed in order to be considered a specific hit, which represents a high confidence association between a protein query sequence and a conserved domain and therefore a high confidence level for the inferred function of the protein query sequence. The threshold is equal to the weakest E-value (and highest bit score) among self-hits of a domains member protein sequences to the resulting domain model (illustrated example). Domain-specific threshold scores are calculated only for NCBI-curated domains."

The code in makedatabase.py was modified accordingly:

    print('* Making peptide-specific reference database...')
    with TemporaryFilePath() as tempfasta, TemporaryFilePath() as tempfasta2:
        filterFASTAByHMM(
            hmm_model=args.hmm,
            input_fasta=args.data,
            output_fasta=tempfasta,
            hmmer_output=hmmer_output,
            additional_args=f'--cut_tc'
        )