🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

antisense. Omar Wagih bioRxiv bioRxiv and medRxiv are undergoing a platform upgrade that has led to unanticipated performance issues. Apologies for the inconvenience - these should be resolved soon.

Single cell foundation models have the potential to revolutionize many areas of biology. How does the composition of training data affect the performance of these models? In my intern project Microsoft Research New England, we identify patterns that can augment next-generation models.

presenting CleaveNet: an AI pipeline for the generative design of protease substrates We designed novel, selective substrates for MMPs by conditioning on an enzyme activity profile and validated them through a large-scale in vitro screen. for protease biology and beyond! 🚀

🚨Online now! 📄Artificial variables help to avoid over-clustering in single-cell RNA sequencing 🧑‍🤝‍🧑 @alandenadel & colleagues cell.com/ajhg/abstract/…

Great to see our paper presenting recall, a framework which calibrates clustering for the impact of data "double-dipping" in single-cell studies, out today in AJHG! Congratulations, Alan DenAdel and co-authors!

Over-clustering of single-cell RNA-seq data can produce spurious results. Alan DenAdel, Lorin Crawford, & co of AJHG latest study introduce recall, a new method, protecting against over-clustering & enables rapid analysis of single-cell RNA-seq data: cell.com/ajhg/abstract/…

Our paper has been published Nature Genetics! Through new statistical methods, we shed light on fundamental questions about cellular response to genetic perturbations. Our work is a substantial advance towards rigorous characterization and comparison of massive perturbation atlases.

Excited to share our latest preprint, introducing the hierarchical cross-entropy (HCE) loss — a simple change that consistently improves performance in atlas-scale cell type annotation models. doi.org/10.1101/2025.0…

How well do single-cell foundation models perform w/o finetuning? Our work Genome Biology shows that in zero-shot settings, scGPT and Geneformer often underperform traditional methods, raising questions about their utility for biological discovery. 📰 genomebiology.biomedcentral.com/articles/10.11…

Defining “what is good” is central to both AI and Science. Despite massive efforts towards building the Virtual Cell, our recent Nature Biotechnology article reflects on whether current gold standards (i.e., evaluation metrics) still hold up. nature.com/articles/s4158…

In 1965, Margaret Dayhoff published the Atlas of Protein Sequence and Structure, which collated the 65 proteins whose amino acid sequences were then known. Inspired by that Atlas, today we are releasing the Dayhoff Atlas of protein sequence data and protein language models.

thrilled to share The Dayhoff Atlas of protein language data and models 🚀 protein biology in the age of AI! aka.ms/dayhoff/prepri… we built + open source the largest natural protein dataset, w/ 3.3 billion seqs & a first-in-class dataset of structure-based synthetic proteins