iiinova — where your digital self keeps evolving

briantrippe.bsky.social

@briantrippe.bsky.social

150 followers 137 following 8 posts

Assistant Professor at Stanford Statistics and Stanford Data Science | Previously postdoc at UW Institute for Protein Design and Columbia. PhD from MIT.

Posts Replies Media

Reposted by @briantrippe.bsky.social

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 19/02/2025

With Zhuoqi Zheng, Bo Zhang, @kdidi.bsky.social @jasonyim.bsky.social @josephwatson.bsky.social Hai-Feng Chen, @briantrippe.bsky.social

021

briantrippe.bsky.social @briantrippe.bsky.social · 19/02/2025

And thanks as well to ChatGPT for turning my first banal attempt at a tweet-thread into emoji-packed click-bait for real work! 😂

020

briantrippe.bsky.social @briantrippe.bsky.social · 19/02/2025

This has been a big team effort – with contributions across 5 timezones 🌍. Many thanks to my fantastic coauthors on our whitepaper: arxiv.org/abs/2502.12479

120

briantrippe.bsky.social @briantrippe.bsky.social · 19/02/2025

This is a V0 pilot—we need your input for V1! 🧩 Know of an important motif? Add it to the benchmark! 🏗️ Help improve the pipeline & metrics (sequence-based? Side-chain-level?) Let’s shape the future of motif scaffolding together!

100

briantrippe.bsky.social @briantrippe.bsky.social · 19/02/2025

Surprise: A modern baseline (RFdiffusion) fails on motifs scaffolded into de novo enzymes 15 years ago 🤯 This suggests modern deep learning methods aren’t always better than past methods!

121

briantrippe.bsky.social @briantrippe.bsky.social · 19/02/2025

MotifBench fixes this by providing: 🧪 A standardized evaluation pipeline 🏆 30 challenging motifs as test cases 📊Easy-to-use eval scripts and a leaderboard for method comparison Now, results can be easily and consistently measured.

100

briantrippe.bsky.social @briantrippe.bsky.social · 19/02/2025

Recent progress in motif scaffolding has been exciting! But… ❌ Current evaluation are inconsistent, and results incomparable ❌ Widely used test cases are too easy ❌ Reproducibility is difficult That’s where MotifBench comes in.

100

briantrippe.bsky.social @briantrippe.bsky.social · 19/02/2025

Motif scaffolding is a core challenge in protein design: ✅ Input: a motif (small functional substructure) 🎯 Goal: find a scaffolds (full proteins) that preserves the motif’s geometry. But what's the state of methods for this problem?

110

briantrippe.bsky.social @briantrippe.bsky.social · 19/02/2025

🔥 Benchmark Alert! MotifBench sets a new standard for evaluating protein design methods in motif scaffolding. Why does this matter? Reproducibility & fair comparison have been lacking—until now. Paper: arxiv.org/abs/2502.12479 | Repo: github.com/blt2114/Moti... A thread ⬇️

14117