comBGC column BGC_length inaccurate

nf-core / funcscan

(Meta-)genome screening for functional and natural product gene sequences

https://nf-co.re/funcscan

MIT License

69 stars 20 forks source link

comBGC column BGC_length inaccurate #399

Closed martinklapper closed 2 months ago

martinklapper commented 2 months ago

Description of the bug

The BGC_length column often reports the BGC length with 1 nucleotide more than the actual contig length, which is not possible. The calculation of this column seems to be wrong. bug_BGC_length.xlsx

Command used and terminal output

No response

Relevant files

No response

System information

No response

jfy133 commented 2 months ago

@jasmezz is this already fixed in the pipeline version V2?

jasmezz commented 2 months ago

Fixed it for antismash on a new branch (#402). The one-based counting in Genbank files vs. zero based in biopython confused me apparently... (see here).
GECCO is correct
DeepBGC reports it in its output, so comBGC just takes it from there. I am not 100% sure how DeepBGC calculates the positions/length. Will take a closer look when time, if it's not correct, need to fix on their end.

jasmezz commented 2 months ago

Done in #402