C2Bio

Curated weekly digest of notable AI/ML and biotech developments. Flagged items are manually selected; remaining items are the week's most recent. 🧬 Biotech ★ ViMax: Agentic Video Generation — Long-form video generation requires systematic narrative planning and visual consistency that current short-clip methods cannot provide. ★ Decoding intercellular activities for more than 1,000 secreted proteins ★ A light-controlled tetanus neurotoxin enables rapid and reversible synaptic inhibition in vivo ★ Alignment with experimental data improves protein generative modeling ★ Efficient evidence-based genome annotation with EviAnn — Machine-learning-based ab initio gene finders have long been central to eukaryotic genome annotation, largely because gene expression data were historically expensive and limited. ★ ClairS: a deep-learning method for long-read tumor–normal pair somatic small variant calling — Somatic variant discovery in tumors is crucial for clinical ana...

Curated weekly digest of notable AI/ML and biotech developments. Flagged items are manually selected; remaining items are the week's most recent. 🧬 Biotech ★ CAFE: A Co-folding Approach for Fragment Exploration of Allosteric and Cryptic Binding Sites — Co-folding models hold immense potential for allosteric drug discovery, but have been severely hampered by their systematic bias toward orthosteric ligand binding. ★ Decoding cancer circulating transcriptomic signatures with language models — Current liquid biopsy methods for multi-cancer detection using plasma cell-free RNA (cfRNA, short RNA fragments circulating in blood that can reflect disease states) typically rely on gene annotations, which can overlook signals from unannotated or repetitive genomic regions. ★ An expanded codebook of human transcription factor DNA-binding specificity — Abstract Gene expression is regulated by transcription factors (TFs), which recognize specific DNA sequence motifs. ★ Generat...

Curated weekly digest of notable AI/ML and biotech developments. Flagged items are manually selected; remaining items are the week's most recent. 🧬 Biotech ★ A systematic evaluation framework for universal antibody-antigen binding affinity prediction and candidate recommendation ★ Miniaturizing and modifying natural proteins with Raygun ★ Cellular and signalling mechanisms that regulate the blood–brain barrier — The blood-brain barrier (BBB) is a unique specialization of central nervous system (CNS) capillary endothelial cells that controls molecular traffic between the blood and the brain. ★ Structure-free, site-resolved contrastive learningextends small-molecule discovery beyond the reachof structure-based modeling — Virtual screening asks which molecules, among an enormous space of drug-like chemistry, are worth synthesizing and testing against a protein target. ★ Structure and evolution-guided design of minimal RNA-guided nucleases — The design of RNA-guid...

Curated weekly digest of notable AI/ML and biotech developments. Flagged items are manually selected; remaining items are the week's most recent. 🧬 Biotech ★ AI-redesigned starting points and outcomes enhance protein evolution — Abstract Engineered or laboratory-evolved proteins often have suboptimal stability, activity or specificity. ★ Distributional regression using generalized additive models for location, scale and shape ★ High-throughput machine learning-aided antibody discovery for cell surface antigens — Machine learning (ML) has the potential to revolutionize antibody design and selection, but its success depends on access to well-curated datasets of antibody-antigen interactions.

Curated weekly digest of notable AI/ML and biotech developments. Flagged items are manually selected; remaining items are the week's most recent. 🧬 Biotech ★ Biotech's digital twin dilemma — Nature Biotechnology explores how digital twins — from virtual patients to simulated cells and organs-on-chips — are forecasting disease and guiding drug development. ★ Universal cell embedding provides a foundation model for cell biology — A self-supervised single-cell foundation model (>650M-parameter, 33-layer transformer) trained on 36M cells across eight species, producing a 1,280-dimensional universal embedding that enables zero-shot cross-species cell-type mapping without fine-tuning or annotation. ★ De novo design of orthogonal far-red, orange, and green fluorophore-binding proteins for multiplexed imaging — Fluorescent proteins and small-molecule dyes offer complementary advantages for biological imaging: proteins are amenable to genetic tagging, whereas dyes pr...

Curated weekly digest of notable AI/ML and biotech developments. I really liked the paper " Guiding generative models to uncover diverse and novel crystals via reinforcement learning" by  Hyunsoo Park & Aron Walsh, published in Nature Machine Intelligence. They leveraged RL to guide the generative model to generate novel crystals that differ from the training data distribution, which also showed more stable training behavior. Traditionally model generate new data by sampling in latents space close to the original training data, which limits novelty, and may even hinder the discovery of entirely new structures that are no longer subject to evolutional selection. Very nice work.  Although I do not work on crystal structures, I do feel their work is inspiring for bio-centered AI/ML applications. 🧬 Biotech ★ TranscriptFormer: A generative cell atlas across 1.5 billion years of evolution — A generative foundation model trained on 112 million cells across 12 species ach...