Nature：Machine Learning 主题· Anqi Chen·· 21 小时前AI 评分62
RFdiffusion 从头计算设计金属蛋白酶
Computational design of metalloproteases
AI 导读
一篇 Nature 论文用 RFdiffusion 方法从头设计蛋白质，构建出可激活水的锌基活性位点并邻近肽段底物结合口袋，得到高活性、可编程的蛋白酶。这类金属蛋白酶在治疗、生物技术和生物修复方面有潜在应用。
正文
Nature, Published online: 14 October 2026; doi:10.1038/s41586-026-11167-2
RFdiffusion methods are used to de novo design proteins with water-activating zinc-based active sites adjacent to a peptide substrate binding pocket, yielding highly active, programmable proteases with potential applications in therapeutics, biotechnology and bioremediation.
来源：Nature：Machine Learning 主题 · nature.com