Novo Nordisk taps Anthropic's Claude to assist drug research
Novo Nordisk has partnered with AI startup Anthropic to deploy its Claude model across early drug discovery and research operations.
The Wall Street Journal (subscription required)
News
Vera's atacicept stabilizes kidney function in Phase 3 trial for IgA nephropathy
Phase 3 clinical trial results demonstrate that Vera Therapeutics' dual BAFF and APRIL inhibitor atacicept, branded as Trutakna, successfully preserved kidney function in patients with IgA nephropathy.
GenomeFrontier and Cellares partner on automated CAR T cell manufacturing
GenomeFrontier Therapeutics will adapt the production of GF-CART01, an autologous virus-free CAR T-cell therapy, to run on Cellares' automated Cell Shuttle system.
GEN, also reported by Fierce PharmaEndpoints News (subscription required)Skyhawk reports 15-month clinical benefit for oral Huntington's drug
Skyhawk Therapeutics reported that patients receiving SKY-0515 showed statistically significant clinical improvements over a control group from nine months through 15 months of treatment in an ongoing mid-stage trial.
Novo Nordisk taps Anthropic to integrate Claude into drug research
Novo Nordisk plans to deploy Anthropic's Claude Science model to assist with scientific reasoning across drug research and development and agentic software tasks.
Reuters, also reported by BioPharma DiveThe Wall Street Journal (subscription required)Novartis exercises $125 million option on Sironax brain-delivery platform
Novartis triggered its acquisition option 14 months after entering an initial evaluation agreement, gaining control of technology engineered to transport large-molecule therapies across the blood-brain barrier.
Fierce Biotech, also reported by BioPharma DiveSling secures $123 million to advance oral thyroid eye disease drug into Phase 3
Forbion led the Series C financing alongside TPG Life Sciences Innovations and Sectoral Asset Management to fund a 120-patient trial of linsitinib.
Fierce Biotech, also reported by Endpoints News (subscription required)BioPharma DiveNovartis drops TREM2 activator in ALS after Phase 2 failure
Novartis halted development of VHB937 in amyotrophic lateral sclerosis after the TREM2 stabilizer and activator missed both primary and secondary endpoints in a 251-patient Phase 2 trial against placebo.
Fierce Biotech, also reported by Endpoints News (subscription required)ByteDance spins out AI drug unit with $290 million financing
Chinese internet company ByteDance completed a $290 million funding round for its AI-driven drug discovery division following its spin-out as an independent company.
Publications
Machine learning picks out tumor-targeting immune cells from single-cell data
A machine-learning algorithm called PreGame identifies tumor-reactive gamma delta T cells from single-cell surface-protein and RNA sequencing data (CITE-seq), separating functional killers from bystander immune cells.
Sex-stratified biological aging models track organ decline across 15 body systems
Investigators trained 38 sex-specific biological aging clocks across 15 organ systems to capture physiological divergence masked by standard sex-pooled models.
Protein language model identifies hidden RNA viruses across 10,000 environmental samples
A computational pipeline called Rider combines protein language models with structural validation to identify divergent RNA viruses from more than 10,000 environmental metatranscriptomes.
Preprints
Digital twin model simulates reduced-dose epigenetic drug schedules in breast cancer
Researchers built a mechanistic computational model combining chromatin dynamics, in vitro response data, and drug distribution kinetics to simulate how triple-negative breast cancer responds to combination epigenetic therapy.
Algorithm cleans background DNA noise out of single-cell chromatin profiles
A computational team introduced AmbientMapper, a generative model designed to decontaminate high-throughput single-cell combinatorial indexing experiments by learning ambient background profiles from empty droplets.
Sampling method determines whether AI-designed antibodies succeed in the lab
A wet-lab validation study comparing generative adversarial networks and variational autoencoders found that sampling protocols, rather than model architecture alone, dictate the experimental success of computer-designed antibodies.
Machine-learning potentials compute enzyme reaction barriers without quantum mechanics costs
A computational study demonstrates that combining machine-learned interatomic potentials with electrostatic embedding can accurately calculate catalytic reaction barriers in designed enzymes without costly quantum mechanical simulations.
Gene-disruption maps in human T cells sharpen immune drug target selection
Researchers built IGNITE, a semi-supervised machine learning framework that ranks therapeutic targets by combining human genetic evidence with Perturb-seq profiles from 22 million primary human CD4+ T cells.
Orchestrated small language models answer biological queries on single-cell data
A multi-agent framework called scACORN coordinates specialized small language models rather than a single monolithic model to interpret single-cell transcriptomic profiles.
Structural homology searches show high error rates for uncharacterized protein domains
A retrospective benchmark using Pfam updates revealed that structural homology searches using Foldseek recovered correct functions for only 15.9% of previously uncharacterized protein domains, compared with 30.1% for difficulty-matched known domains.
Bayesian model maps protein regions intolerant to mutation across the genome
A statistical model named PRIME uses a hierarchical Bayesian framework to estimate regional missense intolerance across genes by sharing evolutionary information across sub-genic segments.
High-resolution microscopy and AI map amyloid assemblies in Alzheimer's mouse model
An automated spatial computing pipeline analyzed nanoscale expansion microscopy images of brain tissue from the somatosensory cortex of 5xFAD Alzheimer's disease mice.