Introducing RNAFlow, a protein-conditioned RNA🧬generative model for structure and sequence design. Excited to present this work, mentored by @WengongJin, @icmlconf in a few weeks!. Paper: Thread: ⬇️🧵

RT @LeoTZ03: Finally kicking this off!!! .Very excited about the series, and we have many exciting speakers—come join us on Wednesday!. Upc….

RT @ml4proteins: We are excited to announce the launch of our new series of talks for early career scientists, including assistant professo….

RT @SGRodriques: Today, we’re announcing the first major discovery made by our AI Scientist with the lab in the loop: a promising new treat….

RT @cognition_labs: Introducing Devin 2.0: a new agent-native IDE experience. Generally available today starting at $20. 🧵👇 https://t.co/….

RT @jiarlu: 🤔Excited about Protein + AI? Join our "AI for Protein Design" tutorial at #AAAI2025, with gorgeous @Oxer22 @divnori @WengongJin….

RT @BrianHie: We trained a genomic language model on all observed evolution, which we are calling Evo 2. The model achieves an unprecedent….

RT @ml4proteins: This Tuesday, 1/7 @ 4 pm EST, we'll be back with the 2025 seminar series!. @SeyoneC will present "Mapping the combinatoria….

RT @SGRodriques: FutureHouse is launching an independent postdoctoral fellowship program for exceptional researchers who want to apply our….

RT @cognition_labs: Devin is generally available today! . Just tag Devin to fix frontend bugs, create first-draft PRs for backlog tasks, ma….

RT @OwenDugan: 🚀 Exciting news! We're proud to announce OccamLLM, a novel method for exact language-model arithmetic in a single autoregres….

RT @rishabh16_: 🚨🚨 I'm giving an Oral presentation at #MLCB2024 on RNA-FrameFlow this Friday, 6th September at 1:20pm PDT 🧬. 🔗: https://t.c….

Poster #109 at 1:30 pm today!.

If you’d like to chat about RNAFlow, or ML for biology in general, please reach out! Excited to connect @icmlconf 🧬🤓.

We evaluate whether RNAFlow can design the sequence and structure of an aptamer that binds to GRK2, a target of interest for chronic heart failure. We provide a known binding motif of 4-nucleotides. RNAFlow’s predictions resemble the true aptamer’s sequence and structure!

Further, we model the dynamic nature of RNA. Over the course of flow matching inference, RNAFlow generates a trajectory of structures. The final RNA sequence design is conditioned on the last few inference outputs which approximate a conformational ensemble.

A frozen RosettaFold2NA network folds the sequence into a denoised structure to perform flow matching in structural space. RNAFlow can design RNAs with modest sequence and structure accuracy while maintaining novelty! See the paper for details.

RNAFlow uses inverse folding as a bridge to unlock the generative capability of pre-trained structure prediction models, without fine-tuning. We predict a denoised sequence from a noisy backbone, with protein conditioning (using a modified version of gRNAde from @chaitjo!)

In the protein design field, diffusion and flow matching models that fine-tune pre-trained structure prediction networks have excelled at conditional binder design. But this is computationally expensive and necessitates a separate sequence design step.

RNAs can be engineered to perform versatile functions, particularly through their interactions with specific protein binding partners.