Torsten Seemann

Prokka relies on Prodigal to detect genes/ORFs, and often misses broken genes, or false frameshifted genes due to bad homopolymer issues with 454/ION/Pacbio/Minion assemblies.

@thysd Diamond only provides `blastp` (Prot:Prot) and `blastx` (Prot:DNA) Unfortunately the design of abricate needs the query to be the contigs, i need `tblastn` (DNA:Prot) I don't think Abricate is...

To be able to do this, I need to know which database sequences were actually CDS or ORFs. Most of them should be, but we don't always know what frame...

I would like Barrnap to work like RNAmmer in that situation. I think I naively assumed NHMMER would maybe align across it, but it probably doesn't and gives 2 separate...

``` rRNA_205522_53445-53560_DIR- dna 116 rRNA_205522_53810-56760_DIR- dna 2951 rRNA_205522_59270-60682_DIR- dna 1413 ``` ``` % barrnap --reject 0.01 issue01.fna [barrnap] Found: 16S_rRNA rRNA_205522_59270-60682_DIR- L=764/1585 1..764 + 16S ribosomal RNA (partial) [barrnap] Found:...

I don't actually remember changing this, but I suspect it was because abricate is just a glorified BLASTN wrapper and the bundled databases go beyond AMR genes now.

abricate uses DNA:DNA matching resfinder uses PROT:DNA matching could this account for the difference?

Yes, both bacteria and archaea share a 16S model from RFAM. ``` NAME 16S_rRNA ACC RF00177 ``` Barrnap is designed for bacterial isolates. It was not designed to predict kingdom...

Yes, you are right that I do not keep the ARO numbers from the CARD input file. I think I will add a new column called `DATABASE_ACCESSION` for this purpose.

@pepijnhuizinga Would `mlst` be a better tool? Can you provide a link to the Warwick plasmidMLST download page? I think the problem is that I use `-culling_limit` from the BLAST...