Question: antismash summary of many genomes about antismash HOT 2 CLOSED

Hi @alienzj,

I am wondering how you dealt with this question?
My solution so far has been to parse the main gbk file, e.g.

$ less ERR260137_bin.12.p.gbk|grep -A 10 'region '|grep '/product'|sed 's/^.*="//g'|sed 's/"//g'
cyclic-lactone-autoinducer
cyclic-lactone-autoinducer
RRE-containing

This extracts the main result types as shown by the different regions of the index.html output file.
I then repeat this for all my samples and create a presence/absence matrix to compare multiple genomes.

On a separate note, I am also wondering what parameters you have used for processing a large number of genomes in parallel on the cluster? My jobs are currently running in series on a single node using 76 cores and the following parameters:

antismash -v --asf --pfam2go --smcog-trees --tigrfam --rre --cc-mibig --cb-general --cb-subclusters --cb-knownclusters --enable-genefunctions --genefinding-tool prodigal -c 76

However, I suspect that for analyzing on the order of thousands of genomes it will be necessary to submit a job for each genome in parallel and reduce the number of cores (~4 cores/job?).

Please let me know if you can advise me based on your experience with antismash.

Best wishes,
Francisco

from antismash.

@franciscozorrilla
You can refer to https://github.com/ohmeta/oral-assembly/blob/master/notebook/antismash.ipynb for processing antismash''s output, but I am not sure what I did is right.

Based on my experience, if you need to process a large number of genomes, it is recommended to deliver tasks to the cluster to run, like SGE or SLURM.

from antismash.