3 ms·
As an FYI, I've been interested in a similar approach. So far bioinformatics techniques seem to be the best fit. You may want to take a look at libraries that
by ABCLAW 10y ago
As an FYI, I've been interested in a similar approach. So far bioinformatics techniques seem to be the best fit.
You may want to take a look at libraries that provide BLAST or Clustal/MSA Analysis capabilities, then tinker a bit. From my experience, wiring those tools to a separate visualization suite provides the best results.