Ever wonder why our HLA specified cancer therapies are only for HLA02:01 thus far? @possuhuanglab presents the scope of the problem at the inaugural @StanfordCancer AI and Cancer Research Symposium 🧬 #AICancerResearch

Excited by the growing interest in Caliby! Based on feedback from a few groups, we've downgraded some dependencies to support older OS versions. Please give it a try, and feel free to reach out with any issues https://t.co/m8bJtlJZKa

A protein platform rapidly develops peptide-focused major histocompatibility complex class I binders with high specificity https://t.co/TWNXuNo0zE https://t.co/xOfPhTboxp

Specific binders to peptide–MHC class II are rapidly generated without laborious screening https://t.co/V7N6AGYPtV https://t.co/UwDJy4S8Sh

Tired of the bottleneck ⏳👔 between screening protein binders and actually measuring their affinities? Then check out our recent pre-print:

Interesting feature of the SLAE latent space ⬇️

Read our preprint here:  https://t.co/BL9pBcPdz9  (8/8)

Work done by @_YilinChen_, @Tianyu_Lu in the @PossuHuangLab, Cizhang Zhao and @HWaymentSteele. Thank you all! (7/8)

SLAE projects all-atom structures onto a smooth manifold! Unguided linear interpolation between conformations in SLAE latent space decodes to coherent intermediates structures. (6/8)

SLAE extends our generative coverage assessment SHAPES to all-atom, per-residue-type granularity. Now we can compare de novo all-atom protein design models and spot residue-level environment biases. (5/8)

Rich in atomic-environment signal, SLAE features outperform PLMs and task-specific models across diverse, challenging downstream tasks, including binding affinity, thermostability and chemical shift prediction. All-atom structure pretraining is all you need! (4/8)

The SLAE latent landscape is organized in meaningful ways beyond amino acid identity. It separates residue embeddings along features including solvent accessibility, secondary structure and structural nativeness. (3/8)

We design a deliberately hard two-part task to learn compact, expressive features. A local graph encoder projects each residue’s atomic interactions into a feature vector, while a global decoder learns to compose these local environment tokens into coherent macromolecules. (2/8)

Introducing SLAE, our new framework to represent all-atom protein structures with residue local chemical environment tokens! SLAE reasons over atomic interactions to recover full structures and residue pairwise energetics, yielding a generalizable latent space. (1/8)

Excited to share Caliby 🐈, our new model for structure-conditioned sequence design! Caliby is a Potts model-based sequence design method that can condition on structural ensembles. We use this to average out non-structural signal (e.g. evolutionary bias) learned by models 🧵1/N

💻Sampling and training code for Protpardelle-1c is now available: https://t.co/qA3x4jfiw7 Feedback and requests are welcome!

We were invited to write a preview about SHAPES, a great recent work from @PossuHuangLab. I really enjoyed this paper! It shows how far we still are from sampling the protein structural space without bias. Our preview just came out, check it out here: https://t.co/Pyy0AFpeI2

Code will be released soon on our GitHub: https://t.co/ei8EyBIuQU Preprint: https://t.co/z2pVpCzkif Have fun sampling and training!

Our new set of all-atom models can sample plausible sidechains without stage-2 sampling. Sequence-dependent partial diffusion behavior occurs when we mask the dummy atoms.

https://t.co/bW5dXvo2FJ