diamond icon indicating copy to clipboard operation
diamond copied to clipboard

--cluster-steps parameter set as 'sensitive' in clustering

Open AMbioinformatics opened this issue 2 years ago • 6 comments

I try to run the following comand: diamond cluster -d INPUT_FILE -o OUTPUT_FILE --cluster-steps sensitive --member-cover 80 -e 1e-05, but after a while, a memory-related error appears, even though the memory is set to 300 GB.

When I run the following command: diamond cluster -d INPUT_FILE -o OUTPUT_FILE --member-cover 80 -e 1e-05 everything is fine.

I would like to set parameter --cluster-steps as 'sensitive'. Am I doing this in the correct way? What could be the cause of the error?

AMbioinformatics avatar Dec 28 '23 18:12 AMbioinformatics

How big is your input file? Directly going to --sensitive is very expensive for larger files.

bbuchfink avatar Jan 08 '24 14:01 bbuchfink

@bbuchfink 12 GB

AMbioinformatics avatar Jan 09 '24 18:01 AMbioinformatics

I'm not sure why this would run out of memory. In any case, sensitive all-vs-all of a file that size will be expensive. Normally you would use cascaded clustering e.g. --cluster-steps faster_lin fast default sensitive. Or do you specifically intend not to use cascaded clustering?

bbuchfink avatar Jan 11 '24 10:01 bbuchfink