iiinova — where your digital self keeps evolving

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 05/10/2026

Combine a chemical knowledge, generative modeling, and structure prediction into a pipeline that can design enzymes for new-to-nature chemistry. @zacharywu.bsky.social @alexechu.bsky.social @francesarnold.bsky.social @pushmeet.bsky.social @juewang.bsky.social www.biorxiv.org/content/10.6...

0124

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 23/09/2026

A genome language model reveals the interactions hidden in non-coding DNA, pointing to new putative RNAs and repetitive elements in bacterial genomes. @garykbrixi.bsky.social @maxewilkinson.bsky.social @mfgrp.bsky.social @brianhie.bsky.social www.biorxiv.org/content/10.6...

1177

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 22/09/2026

Use reward-guided search in a generative model's latent space to design high-affinity protein binders to proteins and carbohydrates. @karstenkreis.bsky.social @pierceogden.bsky.social www.biorxiv.org/content/10.6...

0157

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 18/09/2026

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...

0111

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 16/09/2026

FoldDir: A flow-matching structure-conditioned protein sequence design model with excellent computational results and some functional nanobody redesigns. @rbaltman.bsky.social www.biorxiv.org/content/10.6...

072

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 14/09/2026

A nice review on ways to discover and engineer proteins that use fewer amino acids while maintaining structure and function. @FilipBuchel @KlaraH_lab www.sciencedirect.com/science/arti...

051

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 10/09/2026

The cost of gene synthesis prevents machine-learning guided directed evolution from reaching it's full potential. A very nice perspective by the inimitable Bruce Wittmann. arxiv.org/abs/2609.03046

072

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 08/09/2026

A fast method for building phylogenetic trees with millions of sequences and then a model that leverages the tree structure to predict viral protein evolution! @seyonec.bsky.social @claudiadriscoll.bsky.social @garykbrixi.bsky.social @brianhie.bsky.social www.biorxiv.org/content/10.6...

0153

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 31/08/2026

The largest experimental protein sequence-fitness dataset: 13k measurements across 3 proteins reveals that fitness landscapes are rugged, there are many functional sequences between orthologs, and ML models can't distinguish functional and non-functional sequences. www.biorxiv.org/content/10.6...

0174

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 28/08/2026

Use machine learning guided directed evolution to make a red-shifted fluorescent protein 4x brighter. www.biorxiv.org/content/10.6...

0101

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 24/08/2026

Curate a molecule <-> text dataset, train a model, and use it to discover molecules that fight antibiotic resistance by deactivating beta-lactamase enzymes. @kosonocky.bsky.social www.biorxiv.org/content/10.6...

1132

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 20/08/2026

Binder design is nice and all, but here we have three agents sharing data and autonomously controlling a lab to design enzymes with shifted substrate scopes and high activity! @cobanbrooks.bsky.social @pascalnotin.bsky.social @philromero.bsky.social www.biorxiv.org/content/10.6...

0205

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 12/08/2026

ApexFold: condition predictions of peptide secondary structure on the chemical environment @delafuentelab.bsky.social www.biorxiv.org/content/10.6...

0112

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 10/08/2026

Use machine learning to design libraries of enzymes with broad substrate scope as starting points for engineering. @francesarnold.bsky.social www.biorxiv.org/content/10.6...

1182

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 06/08/2026

Use a rank regression model to engineer improved enzymes www.nature.com/articles/s41...

091

Reposted by Kevin K. Yang 楊凱筌

Luca Pinello @lucapinello.bsky.social · 17/07/2026

1/ I'm excited to share that we're launching NECB 2026, the inaugural New England Computational Biology Symposium. Oct 1-2 at Microsoft Research New England, Cambridge. Two days of keynotes,talks, and posters to bring our community together across institutions.Space is limited. newenglandcompbio.org

11710

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 10/07/2026

Gaussian process kernels that encode evolutionary information and local linearity improve protein sequence-fitness predictions. arxiv.org/abs/2606.11057

091

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 01/07/2026

Proto: A probabilistic programming language for designing biomolecules. Absolutely beautiful paper from top to bottom. @adititm.bsky.social @bviggiano.bsky.social @brianhie.bsky.social www.biorxiv.org/content/10.6...

0102

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 19/06/2026

TadA-Bench: a 31-round, million-variant wet-lab enzyme engineering dataset. arxiv.org/abs/2606.02624

0121

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 15/06/2026

Aggregating many experiments into one zero-shot protein language model score obscures that current models cannot meaningfully rank a set of fit mutations or prioritize new-to-nature functions @clauswilke.com www.biorxiv.org/content/10.6...

0216

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 29/05/2026

Learn which structural domains are compatible with each other, then generate protein sequences conditioned on the desired structural domains. www.biorxiv.org/content/10.6...

070

Reposted by Kevin K. Yang 楊凱筌

Jase Gehring @skyjase.bsky.social · 15/05/2026

very very cool! i had the privilege of asking Mohammed AlQuraishi's opinion on this topic a couple years ago

061

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 15/05/2026

Screen 1M random protein sequences to discover that biology-like folds are accessible from random sequences with surprising frequency www.biorxiv.org/content/10.6...

1368

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 11/05/2026

Genie3: all-atom SE3-equivariance for fast and performant protein design. @yeqinglin.bsky.social @moalquraishi.bsky.social www.biorxiv.org/content/10.6...

0102

Reposted by Kevin K. Yang 楊凱筌

Clay Kosonocky @kosonocky.bsky.social · 21/04/2026

Have you wondered what the wet lab success rates are for current AI-driven protein design models? Look no further! In our new review, @kevinkaichuang.bsky.social @avapamini.bsky.social, @sarahalamdari.bsky.social, and I report wet lab success rates for *over 200* different protein design tasks 🧬💻

13214

Reposted by Kevin K. Yang 楊凱筌

Alex Lu @alexijie.bsky.social · 18/04/2026

Check out how we exploit observations of homology from evolution to design enhancers, even when we don't have prior knowledge or ability to specify function!

032

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 17/04/2026

EnhancAR: Use evolution and deep learning to design enhancers with desired expression profiles! Had a lot of fun working with Andrew Duncan, Micaela Consens @alexijie.bsky.social @lcrawford.bsky.social Jennifer Mitchell and Alan Moses! www.biorxiv.org/content/10.6...

0133

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 14/04/2026

A very nice review of AI in protein engineering from @jlistgarten.bsky.social www.science.org/doi/10.1126/...

0143

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 10/04/2026

DISCO: a steerable, multimodal protein diffusion model that can generate enzymes for new-to-nature chemistry. arxiv.org/abs/2604.05181

2173

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 08/04/2026

Sequence Display generates large-scale protein sequence–activity datasets by simultaneously reading out the sequence of interest and a downstream editing efficiency - perfect for data hungry ML methods! Happy to have played a small part in this!

1114

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 06/04/2026

Design enzymes using a model that generates sequence and structure given the substrate as a ligand. @lileics.bsky.social www.biorxiv.org/content/10.6...

0153

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 03/04/2026

Genomic context and protein sequence can be used together to discover bacterial phage defense systems. @emordret.bsky.social @alexhv.bsky.social @audeber.bsky.social www.science.org/doi/10.1126/...

1111

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 02/04/2026

Distilling the most confident mutation effect scores across different versions of the same protein language model results in a superior sequence-only variant effect predictor. @vntranos.bsky.social www.nature.com/articles/s41...

0133

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 13/03/2026

It's great when a big lab expands a subfigure from your paper into its own paper without any acknowledgment

0101

Reposted by Kevin K. Yang 楊凱筌

Diego del Alamo @delalamo.xyz · 26/02/2026

This plot is quite the indictment of fine-tuned PLMs, showing how performance is entirely data-dependent and, at the upper end of performance, equally achievable with randomized model weights

0132

Reposted by Kevin K. Yang 楊凱筌

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 25/02/2026

We made FLIP2, a protein fitness benchmark spanning seven new datasets, including enzymes, protein-protein interactions, and light-sensitive proteins, as well as splits that measure generalization relevant to real-world protein engineering campaigns.

15215

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 25/02/2026

15215

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 24/02/2026

Train a protein language model to predict one homolog from another given the amount of evolutionary time separating them. @antoinekoehl.bsky.social @junhaobearxiong.bsky.social www.biorxiv.org/content/10.6...

0120

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 04/02/2026

Excellent review on using generative models to design enzymes @noeliaferruz.bsky.social @lassemiddendorf.bsky.social arxiv.org/abs/2602.03779

0122

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 24/01/2026

Hot take: it's 2026 we don't need to be manually doing citations or asking llms to do them overleaf should just autogenerate the citations from dois

2274

Reposted by Kevin K. Yang 楊凱筌

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 16/01/2026

For no reason, I remembered today that I too once got to take a picture holding Nobel Prize that I didn't earn

21219

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 16/01/2026

For no reason, I remembered today that I too once got to take a picture holding Nobel Prize that I didn't earn

21219

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 13/01/2026

EDEN: a family of genomic language models trained on up to 9.7 trillion nucleotides from @basecamp-research.bsky.social's BaseData can design large serine recombinases, bridge recombinases, and antimicrobial peptides. www.biorxiv.org/content/10.6... Happy to have played a small part in this!

0195

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 07/01/2026

A dataset of 40 million protein families and an autoregressive model of protein families. Great to see other protein Atlases popping up after Dayhoff!! @judewells.bsky.social @dmmiller597.bsky.social www.biorxiv.org/content/10.6...

0298

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 16/12/2025

Finetune a codon-level language model with 30k tryptophan synthases, then generate diverse, functional, enzymes with broad substrate scopes. Théophile Lambert @jsunn-y.bsky.social @francesarnold.bsky.social www.biorxiv.org/content/10.1...

0224

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 12/12/2025

Enzyme Engineering Database (EnzEngDB): a platform for sharing and interpreting sequence–function relationships across protein engineering campaigns @francescazfl.bsky.social @jsunn-y.bsky.social @francesarnold.bsky.social @arianemora.bsky.social Paper: doi.org/10.1093/nar/... DB: enzengdb.org

1236

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 10/12/2025

An energy-based model of protein conformational space can be used to predict structure from sequence, sample from the conformational landscape, rank structures, and predict mutation effects. @sokrypton.org www.biorxiv.org/content/10.6...

04914

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 06/12/2025

Becoming a real Asian by making my kid practice Chinese characters and math while he waits for his Saturday cello class.

0110

Kevin K. Yang 楊凱筌 @kevinkaichuang.bsky.social · 05/12/2025

Train a model to identify circularly-permuted structural homologs, then use it to discover novel pairs of related proteins! @aidenosinetrip1 @abulnaga.bsky.social @sokrypton.org www.biorxiv.org/content/10.1...

0203