{ "version": "1.0", "report_type": "AIxBio-Daily", "date": "2026-10-16", "title": "AI × Bio 每日简报 · 2026-10-16", "summary": "首期精选近72小时内发布的一手来源：科研智能体流程、开放结构盲测与AI靶点发现合作。以下区分已核对的公告事实、尚未完成的验证及编辑提出的研究机会；不将预测、潜在里程碑或挑战开放误作实验成果、实收收入或性能结果。", "status": "published", "articles": [ { "id": "20261016-001", "event_date": "2026-10-15", "title": "Antigravity展示抗微生物候选检索流程，湿实验验证仍在进行", "category": "科研智能体", "summary": "Google DeepMind展示了由研究者引导的跨数据库、文献与结构证据整合流程。文中候选正进行湿实验验证，当前不是新抗生素疗效的确认。", "technology_path": [ "ai-science", "ai-biology", "life-science-evidence", "scientific-workflow-orchestration", "agent-assisted-evidence-synthesis", "antigravity-science-orchestration" ], "related_technology_ids": [ "antigravity-science-workflows" ], "technical_details": { "problem": "生物学证据分散在多个数据库与分析工具中，研究者需要反复整理和审查。", "method": "智能体协助协调检索、结构查看与证据归纳，研究者持续决定问题与检验方向。", "innovation": "将信息协调与共享三维查看整合为人机协作流程；这是官方展示的流程集成，非通用性能提升证明。", "limitations": "预测功能与机制仍是假设；原文明示候选待湿实验验证，尚不能据此确认抗菌疗效或安全性。" }, "opportunities": [], "sources": [ { "title": "Accelerating the search for antimicrobial peptides with Google Antigravity", "url": "https://deepmind.google/blog/accelerating-antimicrobial-discovery/", "publisher": "Google DeepMind", "type": "official", "published_at": "2026-10-15", "checked_at": "2026-10-16", "verification_status": "verified", "verification_note": "核对页头日期及 Collaborative intelligence、3D canvas 和结尾限制说明；原文明确候选仍在湿实验验证。另读官方科学平台 https://antigravity.google/use-cases/science 与其链接的 science-skills 仓库确认工具定位；平台页无明确发布日期，未用核验日替代发布日期。本条核验的是官方流程描述，不代表独立复现或抗菌疗效确认。", "organization_ids": [ "google-deepmind" ] } ], "verification_status": "verified", "verification_note": "核对10月15日原文日期、研究者引导的证据整合流程与湿实验验证状态；原文尚未报告该候选的实验疗效。" }, { "id": "20261016-002", "event_date": "2026-10-14", "title": "OpenBind首轮盲测开放：400个化合物检验结构预测", "category": "开放评测", "summary": "OpenBind于10月14日开放Zika病毒NS2B-NS3蛋白酶结构预测盲测，覆盖400个化合物。11月11日为阶段截止、12月16日为最终截止，均为23:59 UTC；本次公告未提供模型性能比较，最终榜单须待挑战结束。", "technology_path": [ "ai-science", "ai-biology", "biomolecular-modeling", "protein-ligand-evaluation", "prospective-blind-benchmarking", "multi-metric-structure-blind-tests" ], "related_technology_ids": [ "openbind-blind-evaluation" ], "technical_details": { "problem": "回顾性测试可能受已有结构信息影响，需要在参考结合结构未公开时检验预测方法。", "method": "参赛者提交蛋白质配体复合物预测；官方联合检查姿态误差、界面准确度和物理合理性，而非只看一个距离指标。", "innovation": "为对接、共折叠及组合流程提供统一的前瞻盲测场景；本轮仍允许使用既有公开靶标结构。", "limitations": "这是单靶标结构评测，不衡量药效或临床价值；本次核验限于官方公告，未独立评估实时榜单或参赛模型成绩。" }, "opportunities": [ { "id": "20261016-002-reproducible-blind-baseline", "title": "编辑机会：建立可复现且记录数据暴露的盲测基线", "kind": "research", "description": "编辑判断：可利用本轮挑战检查结构预测流程的边界，区分公开结构带来的帮助与真正的盲测表现；这不是已取得的研究结果。", "next_step": "先阅读官方规则与数据许可，固定代码、模型和输入版本，记录已有结构及训练数据暴露情况；用官方教程检查提交格式，准备可复现的方法说明后再决定是否参赛。", "technology_ids": [ "openbind-blind-evaluation" ] } ], "sources": [ { "title": "The first OpenBind Blind Challenge is now live!", "url": "https://openbind.uk/news/the-first-openbind-blind-challenge-is-now-live/", "publisher": "OpenBind", "type": "official", "published_at": "2026-10-14", "checked_at": "2026-10-16", "verification_status": "verified", "verification_note": "核对页头2026-10-14、400个化合物、11月11日后阶段榜单与12月16日截止表述；这是挑战开放公告，未给出参赛模型性能结果。", "organization_ids": [ "openbind" ] }, { "title": "Announcing the first OpenBind Blind Challenge: Zika virus NS2B-NS3 protease", "url": "https://openbind.uk/news/announcing-the-first-openbind-blind-challenge-zika-virus-ns2b-ns3-protease-2/", "publisher": "OpenBind", "type": "official", "published_at": "2026-10-14", "checked_at": "2026-10-16", "verification_status": "verified", "verification_note": "当前页面页头标注10月14日，时间表另列9月30日预告、10月14日开放、11月11日23:59 UTC阶段截止、12月16日23:59 UTC最终截止；分别记录，不把预告日期当当前版本发布日期。核对数据规模、参考结构暂不开放及三项联合评测条件；另读官方Concept页确认开放数据与盲测定位，无独立成绩可核验。", "organization_ids": [ "openbind" ] } ], "verification_status": "verified", "verification_note": "两篇当前公告均标注10月14日；长公告时间表列9月30日预告，与开放日区分。逐项核对400个化合物、两阶段UTC截止和盲测定义；规则允许比赛期间展示实时榜单，本条不据公告推断其当前成绩。" }, { "id": "20261016-003", "event_date": "2026-10-15", "title": "insitro与AstraZeneca合作AI靶点发现，约定4000万美元首付", "category": "AI药物发现合作", "summary": "insitro宣布将Virtual Human平台用于MASH与心肌病的靶点发现合作。协议约定4000万美元首付，另有最高约14.6亿美元条件里程碑；潜在金额不等于已到账收入。", "technology_path": [ "ai-science", "ai-biology", "human-disease-biology", "causal-target-discovery", "multimodal-genetic-integration", "virtual-human-causal-modeling" ], "related_technology_ids": [ "insitro-virtual-human" ], "technical_details": { "problem": "同一诊断标签下存在不同的疾病生物学，靶点选择需要更细的证据。", "method": "按公告，insitro应用Virtual Human识别疾病驱动机制并推进指定靶点；AstraZeneca负责临床开发、监管和商业化。", "innovation": "将平台能力用于两类疾病的独家研究合作，属于研发与商业安排进展，并非新模型架构发布。", "limitations": "公告未公开完整架构或本次合作的临床验证；潜在里程碑须满足条件，不能当实收金额或药物成功证明。" }, "opportunities": [], "sources": [ { "title": "insitro Announces a Strategic Collaboration with AstraZeneca for AI-enabled Discovery of Novel Therapeutics for MASH and Cardiomyopathies", "url": "https://www.insitro.com/news/insitro-announces-a-strategic-collaboration-with-astrazeneca-for-ai-enabled-discovery-of-novel-therapeutics-for-mash-and-cardiomyopathies/", "publisher": "insitro", "type": "official", "published_at": "2026-10-15", "checked_at": "2026-10-16", "verification_status": "verified", "verification_note": "核对10月15日公告的适应证、双方职责和协议条款：4000万美元约定首付，最高约14.6亿美元为潜在发现/开发/商业里程碑，不能视为已到账收入。公告未提供完整模型架构、目标名单或该合作的临床验证；公司陈述并非独立疗效证据。", "organization_ids": [ "insitro" ] } ], "verification_status": "verified", "verification_note": "独立打开10月15日公司原公告，核对适应证、双方职责、4000万美元首付和最高约14.6亿美元潜在里程碑。来源为合作方公告；未据此声称已到账、靶点疗效或临床成功。" } ] }