
@LilaSciences working on scientific reasoning. prev Prescient Design @genentech MSFTResearch; 🎓PhD @Stanford, SB @MIT
Bay Area
Joined January 2020
- Tweets35
- Following408
- Followers638
- Likes620
I'm excited to share that I've joined @LilaSciences to work on LLM reasoning for protein design and autonomous wet lab discovery! I really think this is one of the most exciting directions in AI for science, and am thrilled to be working with such a great team.
I'm grateful for the three years at Prescient Design @genentech, where I built ML methods that were integrated in drug discovery pipelines. I also had the privilege of working alongside some of the most thoughtful and generous people I've met, many of whom are now close friends!
14
2,024
come say hi today at the @gembioworkshop from 4:20pm-5:20pm, hall 4!
paper link: openreview.net/forum?id=oOBM…
workshop: iclr.cc/virtual/2025/worksho…
📢 Excited to present our poster at the #ICLR2025 @gembioworkshop! I'll be introducing TherAbDesign - a novel, sequence-based framework that efficiently optimizes antibodies toward therapeutically relevant biophysical properties. [1/4]
7
1,954
📢 Excited to present our poster at the #ICLR2025 @gembioworkshop! I'll be introducing TherAbDesign - a novel, sequence-based framework that efficiently optimizes antibodies toward therapeutically relevant biophysical properties. [1/4]
💡Our Solution: We evaluate biophysical rules and use functionally informative definitions as optimization objectives in TherAbDesign, enabling higher throughput screening and smarter candidate design. [3/4]
1
2
611
Amy Wang retweeted
[1/n] Does AlphaFold3 "know" biophysics and the physics of protein folding? Are protein language models (pLMs) learning coevolutionary patterns? You can try to guess the answer to these questions using mechanistic interpretability. But the thing is, more often than not, we know beforehand what we want/expect our model to learn, especially in very well-studied domains like Biology. So, instead of playing the guessing game, we trained a model that explicitly learns different biophysical and biochemical concepts that biologists care about.
In our concept bottleneck protein language model paper, we show that we can train the model with billions of parameters, with interpretability constraints, without performance degradation.🧵
Joint work with @nc_frey @tuomasoi @amywang01 @juliusadml @samuel_stanton_ @taylor_joren @0x00B1 @hcorrada @kchonyc and Joseph Kleinhenz at @PrescientDesign
Amy Wang retweeted
1/🧬 Excited to share PLAID, our new approach for co-generating sequence and all-atom protein structures by sampling from the latent space of ESMFold. This requires only sequences during training, which unlocks more data and annotations:
bit.ly/plaid-proteins
🧵
Amy Wang retweeted
As part of our commitment to open science, I'm excited to share that an alpha version of our protein design code is available!
Check out the tutorials to learn how to designs proteins yourself with guided discrete diffusion, just like the pros
github.com/prescient-design/…
Amy Wang retweeted
I am hiring a #machinelearning scientist on my team @Prescient @genentech please apply if interested in geometric deep learning, generative modeling, de novo protein engineering, and want to be on the interface of structure/ML/experiments/engineering
roche.wd3.myworkdayjobs.com/…
Super excited to be at NeurIPS tomorrow and Sat (12/2-12/3)! Would love to meet up and chat about biophysics+ML 🥳
1
9
I'll be presenting a poster on learning from physics-based features to improve protein property prediction at @workshopmlsb. This is work from an MSR internship, where I got to work with amazing people like @avapamini @alexijielu and @KevinKaichuang!
Amy Wang retweeted
#MotorsInQuarantine will feature @amywang01 from @Dunn_Lab and @MarkDodding. Join us tomorrow to discuss how actin links to cell-cell adhesions and kinesin-1 is folded into the lambda particle when not busy transporting cargo along microtubules. mechanochemistry.org/whatson…
Amy Wang retweeted
New paper! @amywang01 describes the probable mechanism by which alpha-catenin, and likely related proteins, makes a force-sensitive catch bond with F-actin. Also, intriguing clues about the origins of cooperative actin binding. Congrats Amy!
elifesciences.org/articles/8…
How do cadherin-catenin complexes interact cooperatively with F-actin to assemble into large, organized clusters observed in cells? Our new pre-print may provide some insight. Big shoutout to my co-first author @bax1337 for working with me on this! biorxiv.org/content/10.1101/… [1/3]
This directionally dependent cooperativity may contribute towards the polarized organization of actin bundles that regulate morphogenesis, wound healing, and muscle contraction. [3/3]
Read about how interactions between cadherin-catenin complexes and F-actin are strengthened under load: nitter.cf/amywang01/status/15237……
There are some interesting mechanistic hypotheses about cooperative interactions in this pre-print too!
So excited to share a preprint of my first, first-author paper! biorxiv.org/content/10.1101/… [1/4]
5
So excited to share a preprint of my first, first-author paper! biorxiv.org/content/10.1101/… [1/4]
In addition, we show that the other alpha-catenin domains negatively affect F-actin binding under load for cooperatively bound complexes. These findings may have interesting implications for how cytoskeletal linkages are dynamically linked in epithelial tissues. [4/4]
1
3
See my other thread 🧵
nitter.cf/amywang01/status/15237… for a related pre-print that @bax1337 and I posted this week on cooperative cadherin-catenin interactions! It's been a busy week for the @Dunn_Lab and Weis Lab 🎉
How do cadherin-catenin complexes interact cooperatively with F-actin to assemble into large, organized clusters observed in cells? Our new pre-print may provide some insight. Big shoutout to my co-first author @bax1337 for working with me on this! biorxiv.org/content/10.1101/… [1/3]
4
















