shoaib252.bsky.social
👤 Joined on 27 August 2026
Followers
Following
Total Posts
@shoaib252.bsky.socialExample below: github.com/Shoaib-CS/Ad... Intuitively, it makes sense to me... but will it make a difference out there in tools like the ones @lh3lh3.bsky.social and others are maintaining and improving?
Show PostSomething that I thought was bound to happen is starting to be published in preprints like the one below: using Language Models, either protein or DNA, to aide with the sequence alignment of query sequences.
Show PostPruning the Search, Not the Signal: Adaptive-Banding Needleman-Wunsch via Protein Language Model Confidence [new] Using PLM confidence for adaptive DP corridor, it recovers near-exact alignments while significantly reducing DP cells.
Show PostIt will because it is confidence rather than likelihood-based. But, in most cases, the improvement will be low in terms of quality and low in terms of absolute cost (which is not too high for sequence alignment). Not for query searches. A bigger fish would be to catch non-functional sequences.
Show PostPruning the Search, Not the Signal: Adaptive-Banding Needleman-Wunsch via Protein Language Model Confidence https://www.biorxiv.org/content/10.64898/2026.08.26.747234v1
Show Post