Martin's foldseek has an easy-to-use API that allows us retrieve foldseek results easily.
Why useful: Martin already added the AFDB release for the 200M structures (though, he clustered them at 50% SeqID). Still: this gives MUCH better remote homologs than plain sequence search and currently that is a feature that not many people know of and that not many people use. So we would be among the first to offer such an useful service (of course, Martin's web-service exists but it is not well known and it would be useful addition to what we already have, imho).
Martin's foldseek has an easy-to-use API that allows us retrieve foldseek results easily. Why useful: Martin already added the AFDB release for the 200M structures (though, he clustered them at 50% SeqID). Still: this gives MUCH better remote homologs than plain sequence search and currently that is a feature that not many people know of and that not many people use. So we would be among the first to offer such an useful service (of course, Martin's web-service exists but it is not well known and it would be useful addition to what we already have, imho).
Link to their current web-service: https://search.foldseek.com/search Link to their API documentation: https://search.mmseqs.com/docs/
Same goes for the MSA service, though, in my eyes this is less relevant but we could recycle the way we did it for PP: add some link to an MSA viewer as done in the "About your sequence" field in current PP, e.g., https://alignmentviewer.org/?url=https%3A%2F%2Fpredictprotein.org%2Fapi%2Fppc_fetch%3Faction%3Dget%26hash%3Da9f90a5624fe2894b5e1916e761b904b283dfe6d%26file%3Dquery.mmseqs2AliUref