The front page of AI-driven biology
Monday, October 19, 2026Mon, Oct 19
Updated 12:01 PM PT
M Ventures and Revelation Partners co-led the €120 million ($135 million) round to take NOUS-209 into a global, registration-enabling study for cancer interception in Lynch syndrome carriers, joined by Angelini Industries, the European Investment Bank, Neva SGR, 5AM Ventures, and Andera Partners.
The European Medicines Agency's human medicines committee recommended approval for Eli Lilly's oral GLP-1 drug Foundayo for weight management and type 2 diabetes in adults.
Danaher and Nvidia are collaborating to connect laboratory instrumentation and automation hardware directly with edge computing and biological foundation models.
Novartis has terminated a Phase 2 trial assessing Fabhalta in a serious kidney condition, citing strategic considerations in federal clinical trial records.
The FDA issued an early safety alert for four lots of Edwards Lifesciences' Commander M delivery systems, used with its Sapien M3 transcatheter mitral valve replacement device, after balloon bursts led to one death and one serious injury.
China's Ministry of Commerce has proposed adding 40 technologies spanning artificial intelligence, biomedicine, and photovoltaics to its list of prohibited or restricted exports.
City Therapeutics sold 11.1 million shares at $18 each to raise $200 million in gross proceeds to advance its pipeline of RNA-interference (RNAi) drugs, which turn off disease-linked genes.
Researchers introduced CLIMB and CLIFF, a machine learning workflow that separates bulk RNA sequencing data into distinct cell types and resolves overall drug response down to specific cell populations.
Researchers have developed AmpliconSuite, an open-source software pipeline that reconstructs focal gene amplifications, including extrachromosomal circular DNA (ecDNA) and breakage-fusion-bridge cycles, directly from cancer whole-genome sequencing.
Researchers designed synthetic proteins from scratch that bind nanomolar epitopes on MYC, a notoriously difficult-to-drug cancer-driving transcription factor characterized by disordered loops.
A distributed system deployed thousands of coordinated AI agents to independently analyze facets of a whole-brain Perturb-seq screen, completing in minutes what human researchers take months to evaluate.
Researchers combined gene expression data with dynamic flux balance analysis, a computational method that tracks metabolic reaction rates over time, to simulate tumor growth across diverse human cancer cell lines.
A computational framework named SCOTE calculates pair-wise spatial correlations across all tissue length scales to unify readouts between competing spatial biology technologies.
Mutations in the RNA-splicing gene SF3B1 cause noncoding RNAs to yield novel peptides and proteins, according to a study pairing long-read transcript sequencing with a workflow named ALLR.
Generative AI models that synthesize medical scans often dilute real statistical associations between patient phenotypes and clinical outcomes because they drop crucial biological covariates during image generation.
A gradient-boosted machine learning model trained on structural and energetic descriptors from resolved antibody-envelope complexes predicted broadly neutralizing antibody binding affinities across multiple HIV-1 genetic variants with a test R-squared of 0.658.
Simulations revealed an allosteric pocket formed by the unique N-terminal tail of RhoJ, an enzyme driving tumor blood vessel growth that has long resisted selective drug discovery due to similarity with Cdc42.
Researchers built an automated pipeline using structural modules called Metal Mers, which connect protein backbone geometry to preferred metal coordination sites, to construct artificial metalloenzymes directly from computer models.