I want to look at variations of read identity, as a high-level proxy for population variation, on human microbiome sample profiling using humann v 3.8 (i have been unable to install more recent versions).
There is an obvious parameter to set, --nucleotide-identity-threshold.
However, there is no apparent alteration to the run config regarding nucleotide identity thresholds.
humann -i reads.fq -o humann_default --output-basename default --threads 10 --taxonomic-profile metaphlan.txt
grep 'nucleotide identity threshold' humann_default/humann_default_humann_temp/humann_default.log
nucleotide identity threshold = 0.0
humann -i reads.fq -o humann_90nuclid --output-basename default --threads 10 --taxonomic-profile metaphlan.txt --nucleotide-identity-threshold 90
grep 'nucleotide identity threshold' humann_90nuclid/humann_90nuclid_humann_temp/humann_90nuclid.log
nucleotide identity threshold = 0.0
The resulting outputs also suggest that the parameter did not affect the run - abundances of gene families are exactly identical across all >600K gene families with taxonomic information.
As there are newer versions of the tool, i don’t expect a fix to the version I am using. But please check v4.