a passing thought on contaminated references

[clb21565]

hi there-- hope you're well!

working with genome grist output and realizing that contaminated eukaryotic genomes (where, e.g., deep sequencing of eukaryotic isolates from fecal samples) will likely throw sourmash gather through a loop because the set-cover algorithm will decide to allot a LOT of k-mers there bc there are many/multiple genomes. we are seeing this with some wastewater searches of emerging pathogens (e.g., cyclospora cayatenensis)

not really a sourmash problem, but helpful to consider and report.