Closed sheephorse closed 6 years ago
Hi @sheephorse, what you may want is digital normalization; see
http://ivory.idyll.org/blog/what-is-diginorm.html
Let me know if you have any questions! Although diginorm-specific q should be asked over in github.com/dib-lab/khmer/ instead of sourmash.
I would like to reduce my Illumina metagenome to something approaching unique sequences, i.e. remove sequences that are already represented (though not necessarily exact duplicates that could be removed by dedupe utilities). I want to maintain as much new sequence information as possible, while minimizing the size of the database. I could sort of do this with an assembly and mapping approach, but can I do this at the read level with sourmash?