Kevin K. Yang 楊凱筌

kevinkaichuang.bsky.social

18 September 2026 22:36

Use a genomic language model to optimize a polyketide synthase go produce a non-natural product. @nathanlanclos.bsky.social @ancornman1.bsky.social @kortemmelab.bsky.social @keaslinglab.bsky.social @microyunha.bsky.social www.biorxiv.org/content/10.6...

Tatta Bio

tattabio.bsky.social

02 April 2026 14:27

Most protein-protein interaction tools work on protein pairs. FlashPPI runs at proteome scale and now across two proteomes at once. Upload any two datasets (full genomes, partial genomes, or custom protein sets) and get back a predicted interaction network spanning both.

Alex Crits-Christoph

acritschristoph.bsky.social

05 March 2026 19:36

Really cool! Potentially ignorant question- did you benchmark how PPI prediction correlates with genomic context? E.g., how do known genomically distal PPIs perform compared to syntenic PPIs? I know gLM2 is good at syntenic. I see Fig5, so I know non-sytenic works, but curious if you quantified!

Aude Bernheim

audeber.bsky.social

03 March 2026 16:50

🤩🤩

Tatta Bio

tattabio.bsky.social

03 March 2026 15:18

🧵 FlashPPI (new model) is live in SeqHub. Proteome-wide PPI prediction in minutes, 2,400x faster than the next best sequence-based method, 4x better predictive performance. How it works: 👇