Open xvazquezc opened 2 years ago
Are you wanting these as genes in genbank predicted files - infernal predictions can be pretty slow and am not how universally generic a prediction pipeline can be for these as it will depend on the db. Are you wanting to do this for an animal, plant, or fungus?
But I think you can import more gene features as a gff file you provide I think but not sure about how it works for noncoding gene features.
If you can provide some general GFF files from tools to annotate these I can look at what it would take to have funannotate pass them through.
I'm running Infernal on fungal genomes, mostly following the Rfam recommendations, i.e. using -Z
to calculate e-values, only using CM (--nohmmonly
), and using the gathering threshold (--cut_ga
, curated thresholds). The only parameter I change is using the more strict --default
instead of --rfam
at a cost of running time. e.g.
cmscan -Z ${ZVAL} --cpu $NCPUS \
--cut_ga --default --nohmmonly --fmt 2 \
--tblout rfam.tblout \
--clanin ${RFAMDB}/Rfam.clanin \
${RFAMDB}/Rfam.cm $GENOME > rfam.cmscan
So far, the only clear "errors" occur with rRNA genes which can be called as archaeal instead of eukaryotic or bacterial (mito) or the other way around. Nonetheless, rRNA-specific tools are recommended and e.g. barrnap generates gff3 files by default (that could be another nice addition btw)
Regarding Infernal's output, there is no gff option, but the way is run as recommended generates a tabular format output file (fixed width table). Here there is an example.
@nextgenusfs I can give this a try.
For conversion to gff, see jiffy-infernal-hmmer-scripts, in particular infernal-tblout2gff.pl
Hi there,
I'm annotating a few genomes and I realised that funannotate does not have any integrated tool for annotating ncRNA other then tRNAscan-SE. I will be annotating ncRNA through Infernal, as per the standard ncRNA annotation with Rfam.
I've only seen a brief reference to adding other annotation on #481, where it is referred to manual manipulation of some
.tbl
files (which are not necessarily a nice format). I've seen thatfunannotate annotate
has the--annotations
option but it's not clearly documented what'd be the format or if it'd be valid for genes not yet called.Any details that anyone could provide would be appreciated.
Cheers,