Alphabet 旗下 Isomorphic Labs 完成 50 亿美元融资,用 AI 加速药物发现
Alphabet’s Isomorphic Labs raises $5B to accelerate drug discovery with AI - SiliconANGLE
Isomorphic Labs 宣布完成 50 亿美元融资,由 Greenoaks 和 Sequoia Capital 领投,母公司 Alphabet 通过 GV 和 CapitalG 参与投资。其药物设计平台 IsoDDE 用 AI 加速寻找蛋白结合口袋等环节,并称可跳过 FEP 算法所需的三维蛋白建模步骤、输出质量更高。CEO Demis Hassabis 表示本轮资金将用于构建基础生物学模型、加深系统性疾病理解,并推动首批治疗项目进入临床试验。
正文
Isomorphic Labs Ltd. today announced that it has closed a $5 billion investment led by Greenoaks and Sequoia Capital.
The venture capital firms were joined by more than a half-dozen other backers. The group included Isomorphic parent Alphabet Inc., which made a “significant investment.” The search giant provided the capital through its GV and CapitalG startup funds.
Isomorphic offers a software platform called the Isomorphic Labs Drug Design Engine, or IsoDDE for short. It uses artificial intelligence to speed up several of the tasks involved in developing new medicines. The Alphabet unit says that multiple pharmaceutical companies are already using IsoDDE to power drug discovery programs.
Pathogens such as viruses comprise numerous proteins that each power a different biological process. Drugs work by neutralizing a single, particularly important protein without which the pathogen can’t function. Developing a new drug requires researchers to find a vulnerable protein within a pathogen and then develop a compound that can effectively target it.
The first step of the workflow is mapping out the shape of a target protein. Researchers look for so-called binding pockets, recesses on the protein’s surface to which a drug can be attached. The task is highly complicated because proteins don’t have fixed shapes. Their configuration changes when they interact with other molecules or encounter new environmental conditions, which happens often.
Isomorphic says that IsoDDE speeds up the task of finding binding pockets. When there are no such pockets, it helps researchers develop compounds that cause a target protein to form them.
Once scientists identify a binding pocket, they have to find a way of attaching a drug molecule to it. The pharmaceutical industry searches for such binding mechanisms using a method called FEP. FEP algorithms can be highly hardware-intensive, which makes them costly. Moreover, they often require a detailed, three-dimensional model of a target protein.
The process of creating 3D protein models for FEP algorithms is itself a highly complex task. It requires scientists to freeze a target protein in a crystal and scan it using specialized X-ray equipment.
IsoDDE skips that step of the research workflow. Furthermore, Isomorphic says that the software can generate higher-quality output than FEP algorithms.
“Our drug design engine is generating molecules with unprecedented precision and speed,” said Isomorphicc Chief Executive Demis Hassabis. “This round enables us to accelerate our progress: building foundational biology models, deepening our systemic disease understanding, and driving our first therapeutic programs towards clinical trials.”
Isomorphic added that it plans to build new bioengineering infrastructure and grow its partner ecosystem.
Image: Isomorphic Labs
AI 热点