Protein language models don't scale reliably: bigger doesn't always mean better for embeddings. What if what you need from ESM-15B is already inside ESM-8M? Introducing Reverse Distillation, which flips the usual distillation script: small models anchor and decompose large ones

Excited about this! I'll focus on analyzing immune repertoires w PLMs. Biological systems are inherently multi-scale. With the representational power and speed of PLMs, we can now bridge the molecular and systems scales, to study what makes each of us distinctive, immunity-wise.

AIRR-Community@airr-community.bsky.social · last yr.

The next AIRR-C seminar is scheduled for June 26 at 1600 CET, 1000 ET, 700 PT with Rohit Sing @rohitsingh8080.bsky.social (antibody PLMs) and @joshuamcgrath.bsky.social (immunoglobulin codon mutability). Register here: antibodysociety.org/the-airr-com...

Yes, we have lots of exciting collaborative projects at the interface of computation and biology. Deep expertise in many domains between our labs, so a wonderful and committed training environment!

Introducing KolossuS to address a 50-year old problem: which kinases are active in your pathway of interest? As computational biologists, our work mostly involves post-hoc analysis algorithms. KolossuS is the rare case where a ML model enables entirely new capabilities. 1/

The BEST part: This would not have been possible without the close (and SO FUN!) collaboration of my lab with @rohitsingh8080.bsky.social and the lab of Masashi Yanagisawa.

Over 50 yrs since the discovery of protein kinases, 80% of human kinases still have ≤20 known substrates, and many are “dark.” I'm EXCITED to announce our new work towards solving this- combining (1) deep learning with (2) proximity proteomics in vivo! ➡️ www.biorxiv.org/content/10.1...

We had a great morning session, including a keynote on single cells and long reads @aliciao.bsky.social, and talks on spatial transcriptomics @rohitsingh8080.bsky.social, DNA storage, and TAD inference

Enjoyed reading this one. Very insightful with some clever and intricate analyses. Also very well written.

Rohit Singh@rohitsingh8080.bsky.social · 2y ago

Bio foundation models are great design and engg tools. But can they help decode the fundamental principles of life? We harnessed a single-cell FM for decoding the long-debated relationship between genome arch. and gene coregulation. 1/ Preprint here: www.biorxiv.org/content/10.1...

100% agree with this.

Bio foundation models are great design and engg tools. But can they help decode the fundamental principles of life? We harnessed a single-cell FM for decoding the long-debated relationship between genome arch. and gene coregulation. 1/ Preprint here: www.biorxiv.org/content/10.1...

Tracing the Shared Foundations of Gene Expression and Chromatin Structure

The three-dimensional organization of chromatin into topologically associating domains (TADs) may impact gene regulation by bringing distant genes into contact. However, many questions about TADs' fun...

biorxiv.org

Our work on better ways of aggregating PLM embeddings is now out in Bioinformatics Advances.

Bioinformatics Advances@bioinfoadv.bsky.social · 2y ago

🧬 Now in Bioinformatics Advances: "Aggregating residue-level protein language model embeddings with optimal transport" Explore the full paper: https://doi.org/10.1093/bioadv/vbaf060   #ProteinRepresentation #MachineLearning #Bioinformatics  @rohitsingh8080.bsky.social

🧬 Now in Bioinformatics Advances: "Aggregating residue-level protein language model embeddings with optimal transport" Explore the full paper: https://doi.org/10.1093/bioadv/vbaf060   #ProteinRepresentation #MachineLearning #Bioinformatics  @rohitsingh8080.bsky.social

Well well, timely! 😉 www.cell.com/cell-systems...

Should biology put complexity first?

The dictum “Everything should be made as simple as possible, but no simpler” poses a problem for biology. How simply can it be told without doing damage to its complex nature? The answer might be foun...

cell.com

Looking fwd to presenting our works (Velorama, GrID-Net, plus more) on single-cell regulatory inference. All of this work is based on our framework generalizing Granger causality to directed acyclic graphs. If you're at ENAR 2025, come say hi! We're recruiting! #ENAR2025

We've updated the Raygun preprint with additional validations (more fluorescence assays, biotin ligase reengineering, EGF optimization etc.) Take a look! Here's the preprint: www.biorxiv.org/content/10.1...

Miniaturizing, Modifying, and Magnifying Nature's Proteins with Raygun

Proteins have evolved over billions of years through extensive and coordinated substitutions, insertions and deletions (indels). Computational protein design cannot yet fully mimic nature's ability to...

biorxiv.org

In job interviews and college admissions, we focus too much on what candidates have accomplished, and too little on what obstacles they've overcome. The true measure of people's potential is not the height of the peak they've reached. It’s how far they've climbed to get there.

Delighted to be hosting @jianma.bsky.social as a Distinguished Speaker (Lecturer?) at Duke Cell Biology. Great talk on a variety of foods, from Higashi and SpiceMix to Steamboat! Also, some stuff on 3D genome.

Are you studying a pathway and wish you could perturb it? With catalogs like ZINC and Enamine, billions of small molecule perturbagens are just a click away, but what to order? Enter CoNCISE, our finite scalar quantization-based approach to identifying small molecule binders. 1/

Last month I became re-obsessed with deciphering human biology through the lens of "gene age", that is, when each human gene first appeared in the human genome. This was my analysis of key hemoglobin-related genes.

A nice belated Christmas present--AbMAP is out now in @pnas.org! Great thread from Rohit below 🧵⬇️

Rohit Singh@rohitsingh8080.bsky.social · 2y ago

AbMAP, our work on adapting PLMs for antibodies, is now out in @pnas.org A veritable zoo of PLMs now exists, so why should you care? Well, AbMAP reflects an opinionated design philosophy about antibody representation. And based on your needs, it may be a great fit! 1/ www.pnas.org/doi/10.1073/...

AbMAP, our work on adapting PLMs for antibodies, is now out in @pnas.org A veritable zoo of PLMs now exists, so why should you care? Well, AbMAP reflects an opinionated design philosophy about antibody representation. And based on your needs, it may be a great fit! 1/ www.pnas.org/doi/10.1073/...

Likewise, I’m extremely thankful to David Page and @scottsoderling.bsky.social at Duke, who have recognised that biology and information sciences are increasingly intertwined, with software programmable biology becoming a reality through ML informed by passionate and knowledgable (cell) biologists.

I'm thrilled to join @pranam.bsky.social and @rohitsingh8080.bsky.social. Duke has a unique blend of translational scientists that can bring ideas from the wetlab, to software and ML, back into the wetlab, all in service of characterising, engineering and programming biology!

Super excited to welcome Christian Dallago @machine.learning.bio and Alex Tong @alextong.bsky.social as faculty at Duke and as my neighbors on 4th floor Sands! We're building a fantastic ML+Bio community here at Duke, with @pranam.bsky.social and Phil Romero and more. Trainees, come join us!

Christian Dallago@machine.learning.bio · 2y ago

Two major life updates: - I'm moving to Senior Applied Research Scientist in Digital Biology at NVIDIA (Jan '25) - I'm starting a new lab at Duke as Visiting Assistant Prof (early '25) Both roles focus on tackling hard problems in biological machine learning through collaborative research. Long 🧵

Last month's "Voices" in @cp-cellsystems.bsky.social explored the potential impact of computational protein design. Fantastic range of perspectives-- I was honored to contribute. For all the tremendous translational excitement, I'd argue that the basic science potential may be even bigger.

Amazing AI tool by Rohit Singh's lab at Duke. Shrink or Supersize your protein and retain functions- all based on protein language models.

two boxes of mcdonald 's french fries are being taken out

ALT: two boxes of mcdonald 's french fries are being taken out

media.tenor.com

Rohit Singh@rohitsingh8080.bsky.social · 2y ago

Congratulations to @youngsuko9.bsky.social on their top-10 placement in Adaptyv Bio's EGFR-binder competition! It's an elegant pipeline-- using our protein-design method Raygun to expand and modify the template (EGF), then filtering with ProTrek. foundry.adaptyvbio.com/competition 1/

Congratulations to @youngsuko9.bsky.social on their top-10 placement in Adaptyv Bio's EGFR-binder competition! It's an elegant pipeline-- using our protein-design method Raygun to expand and modify the template (EGF), then filtering with ProTrek. foundry.adaptyvbio.com/competition 1/

🚀 Big updates from our lab! We’re diving into deep learning for protein design and proteomics. For a glimpse of what’s ahead, check out our recent collab with Rohit Singh’s lab, which used AI to shrink proteins! 😱 Read more: www.biorxiv.org/content/10.1... @rohitsingh8080.bsky.social

Miniaturizing, Modifying, and Augmenting Nature’s Proteins with Raygun

Proteins are nature’s versatile nanomachines, but engineering them for enhanced function or novel applications remains challenging. Current methods for protein modification struggle to design sequence...

biorxiv.org