Extract module co-ordinates from antismash output. about antismash HOT 5 CLOSED

The NRPS/PKS domains are annotated as aSDomain features in the genbank outputs, so you can find the aSDomain feature with the matching domain_id from the CDS anntoations and use the protein locations annotated there.

The sec_met_domain annotations aren't handled in the same way, as they only really indicate the presence of at least one domain of that type (e.g. antiSMASH doesn't annotate each individual condensations domain as a unique sec_met_domain in the CDS feature annotations).

from antismash.

We deliberately removed this sort of text output file from antiSMASH 5, as everyone has different needs and we can't satisfy everyone with just one file and maintaining many extra files will inevitably go wrong.

The JSON has the core data from which the locations can be derived, but for your case, you may want to use the annotations from the genbank file instead, as the details you need are reasonably simple.

Here's a python snippet that will print out the equivalent of the table you'd like (you'll want to install biopython with pip install biopython, if not already installed). You can use it as a starting point.

from collections import defaultdict
import json

from Bio import SeqIO

module_counts = defaultdict(int)

for record in SeqIO.parse("your_genbank_file.gbk", "genbank"):
    for feature in record.features:
        if feature.type != "aSModule":
            continue
        locus = feature.qualifiers["locus_tags"][0]
        module_counts[locus] += 1
        columns = [locus, module_counts[locus], feature.location.start, feature.location.end]
        print("\t".join(map(str, columns)))

from antismash.

Thank you for taking the time to sort this out. Indeed many ancillary result files is a recipe for disaster.
I shall implement your suggestion. Closing the issue.

from antismash.

@SJShaw Hi, I'm trying to investigate sequential order of biosynthetic domains from the GBK output from antismash 5.1.2.
Using SeqIO, I extracted domains annotated by "/NRPS_PKS=" features, whereas I failed to resolve how to deal with "/gene_functions="biosynthetic (rule-based-clusters)"". The latter annotations do not include coordinates, and I could not align them with /NRPS_PKS domain information.
Is there any workaround to figure out all annotated domains and their order in each CDS using all "NRPS_PKS", "gene_functions="biosynthetic (rule-based-clusters)"", and "sec_met_domain"?

Thanks,
Taehyung

from antismash.

@SJShaw Thanks for a prompt response. I appreciate it.
I understand what you are saying, but I would like to find a workaround on this matter.
For example, I found putative T3-PKS, which contains PKS_KR domain in one of CDSs. This PKS_KR domain was annotated by both "NRPS_PKS" and "gene_functions="biosynthetic (rule-based-clusters)"" labels . However, there were N and C terminal chalcone synthases in this CDS, annotated only by "gene_functions="biosynthetic (rule-based-clusters)"" label. Although chalcone synthase is crucial for T3-PKS rule-based detection, location of these canonical domains will not be accessible, is that correct?
When I am trying to organize all canonical biosynthetic domains, this makes a problem because I cannot order some domains of NRPS_PKS labels with other domains of "gene_functions="biosynthetic (rule-based-clusters)"" labels. So, I had to dump some canonical domains? Do you have any suggestions to fill up this missing information?

from antismash.