biology

OpenAI, PerturbAI, and Scripps Research used thousands of GPT-Rosalind-powered AI agents to analyze a whole-brain CRISPR atlas of millions of mouse brain cells in just 32 minutes, uncovering new leads on Alzheimer's and Parkinson's disease biology through distributed biological reasoning.

Anthropic and Citrini Research used an internal Claude model to autonomously generate protein minibinders against five scorpion neurotoxins, yielding 17 functional binders from 211 tested, marking a step toward AI-driven antivenom discovery.

Show HN: Grimoire for Biology — a new developer tool or platform aimed at biology, shared with the Hacker News community for feedback.

OpenAI is expanding academic access to GPT-Rosalind through a new Rosalind Academic Plan, offering sponsored access to eligible U.S. university labs to support biological research and scientific discovery.

A free, MIT-licensed open-source tool called Anima Agent skill that turns a single biology prompt into detailed, interactive scientific visualizations—such as switchable mechanisms, inspectable molecular structures, and animated processes—rendered as standalone HTML or SVG. The skill instructs Claude to research the biology before drawing, and includes a reusable rendering toolkit with working examples.

一篇深入浅出的长文，讲述大肠杆菌（E. coli）趋化性——单细胞如何通过 CheY 信号分子、受体复合体与鞭毛马达实现类似嗅觉、记忆与决策的'智能'行为，展示蛋白质结构决定功能等系统生物学核心主题。

Anthropic announced a 'Claude-led' discovery of a novel enzyme system in viral DNA resembling CRISPR gene-editing mechanisms, but scientists questioned the claim, calling the results preliminary, unpeer-reviewed, and possibly duplicating unpublished doctoral research.

科学家首次对世界上现存最古老的陆地动物——约194岁的巨龟Jonathan进行基因组测序，发现其在DNA修复、胰岛素调节、线粒体功能等衰老通路中拥有独特基因变异，相关研究发表于《Science Advances》。

The paper proposes Autoregressive Frontier Expansion, a generative framework that grows tree-like branching structures iteratively, using a flow-matching model parameterized by an SO(2)-equivariant GNN to decide bifurcation versus termination at each frontier step. It is evaluated on cortical neuron and botanical tree morphology generation in unconditional, class-conditioned, and morphology-guided settings, closely matching reference distributions.

Two new papers, one in Nature by Vadim Gladyshev's Harvard team and one in Cell by Juan Carlos Izpisua Belmonte's Altos Labs group, propose a new 'loss of cell identity' model of aging alongside the classic wear-and-tear model, highlighting epigenetic drift and mesenchymal drift as potential targets for anti-aging interventions.

Oxygen-deprived underwater zones, often dismissed as "dead zones," may actually hold clues to the conditions that allowed early life to emerge, according to new research.

Anthropic claims its AI model Claude discovered an enzyme system reminiscent of CRISPR, potentially advancing gene editing, though scientists emphasize the need for further validation.

A groundbreaking discovery related to the origins of life was made in a modest setting, potentially offering new insights into biological beginnings.

C5R has built a research facility that is entirely run by the AI model GPT-6 Astra, enabling autonomous design, execution, and observation of experiments across multiple scientific disciplines.

Scientists discover that postmenopausal ovaries exhibit surprising molecular activity, challenging previous assumptions and indicating potential impacts on overall health and aging.

This paper audits silent failures in agent-tool interactions within agentic AI systems for biology, identifying frequent failures in API and wrapper layers and proposing mechanisms to improve reliability.

This article outlines five millennium-style grand challenges for biology, including demonstrating the origin of life, cryopreserving live mice, reverse translation of peptides, enhancing Rubisco enzymes, and engineering cells with non-standard genetic codes.

Google DeepMind collaborated with ScienceStowers to map over 2,500 regulatory patterns in DNA, identifying sequences that act as biological switches to control gene activity.

Clément Delangue tweeted that risks in biology and cybersecurity are concentrated in powerful labs, and open-source AI is a solution to mitigate these risks.

CellRFT introduces a reinforcement fine-tuning framework that uses biological evaluation as direct training feedback to improve single-cell perturbation models, addressing the mismatch between surrogate training losses and biological evaluation criteria.