新湖畔网 （随信APP) | 今年诺贝尔化学奖颁发给了蛋白质设计和结构预测，AI再次获得胜利。
新湖畔网 （随信APP) | 今年诺贝尔化学奖颁发给了蛋白质设计和结构预测，AI再次获得胜利。
【微信/公众号/视频号/抖音/小红书/快手/bilibili/微博/知乎/今日头条同步报道】
北京时间10月9日下午5点45分，2024诺贝尔化学奖揭晓，瑞典皇家科学院将今年奖项授予了大卫·贝克（David Baker）、德米斯·哈萨比斯（Demis Hassabis）和约翰·M·朱珀（John Michael Jumper），三位科学家因为对蛋白质结构和设计的研究而获奖。
三位科学家将分享1100万瑞典克朗奖金。奖金一半是授予大卫·贝克的，获奖理由是表彰他“对计算蛋白质设计的贡献”（for computational protein design），另一半则由德米斯·哈萨比斯和约翰·M·朱珀平分，他们的获奖理由是“对蛋白质结构预测的贡献”（for protein structure prediction）。
其中生物化学家、计算生物学家大卫·贝克1962年10月6日在西雅图出生，他在1989年获得加利福尼亚大学伯克利分校生物化学博士学位，目前他是美国国家科学院院士、华盛顿大学蛋白质设计研究所所长。
上世纪90年代之前，科学家只能通过对现有蛋白质进行定向改造来为其赋予新功能（比如在化学工业中充当催化剂），大卫·贝克和他的团队选择从零开始设计蛋白质，团队绘制了全新的结构，然后通过自己开发的能预测蛋白质结构的软件Rosetta来计算所需的氨基酸序列，而最终他们成功了。
这标志着蛋白质工程领域实现重大突破。因为这样通过计算机应用和算法来直接设计蛋白质，这摆脱了自然界已有蛋白质的限制，让科学家可以直接“定制”蛋白质，从而在疾病治疗、药物研发、生物学、新材料等各个方面都掌握主动权。
随后大卫·贝克还开源了Rosetta，这进一步推动全球科学家对Rosetta进行持续应用和迭代，这也直接影响了德米斯·哈萨比斯和约翰·M·朱珀后续的工作。
德米斯·哈萨比斯1976年7月27日生于伦敦，他在伦敦大学学院获得认知神经科学博士学位，又在麻省理工学院和哈佛大学进行了神经科学和人工智能研究。2010年他和谢恩·莱格（Shane Legg）、穆斯塔法·苏莱曼（Mustafa Suleyman）一起创立了DeepMind，2014年DeepMind被Google收购后他也来到Google，现在他仍然是DeepMind的CEO。
约翰·M·朱珀则更加年轻，他1985年生于美国阿肯色州小石城，2017年他获得了芝加哥大学博士学位，现在他是DeepMind的董事和高级研究科学家。而以不到40岁的年龄就获得诺贝尔奖，这是一项非常了不起的成就。
延伸阅读：1915年威廉·劳伦斯·布拉格（William Lawrence Bragg）获得诺贝尔物理学奖时只有25岁，他至今也是诺奖自然科学类奖项里最年轻的获奖者。
这两位其实并不是传统意义上的化学或生物领域科学家，DeepMind选择蛋白质结构预测这一研究方向，其实主要是为了验证人工智能的能力（在此之前DeepMind带来的AlphaGo刚刚战胜了李世石和柯洁，深度学习在下围棋这个项目上的探索基本告一段落，蛋白质结构预测成为了下一个课题）。
在德米斯·哈萨比斯团队带来AlphaFold后，蛋白质结构预测的准确率迅速提升，随后约翰·M·朱珀加入团队，他和德米斯·哈萨比斯一起主导了AlphaFold2的开发，这套深度学习算法能“以惊人的准确度”完成对蛋白质的三维结构预测，在事实上终结了过去数十年来科学界在这件事上传统的工作流程，这甚至已经不再是个课题。
AlphaFold2随后也被DeepMind做了开源，它已经被全球相关科研人员应用到了日常的研究当中（需要注意的是这样的预测不是完全准确的，但它能帮助科研人员预测结果的可靠程度。并且过去获得一个蛋白质结构通常要数年时间，还不保证成功，现在用AlphaFold2只需要几分钟就能完成）。
有意思的是，在前一天的物理学奖授予机器学习领域的两位先驱之后，这次化学奖也再次给到两位AI领域的大佬（Google这一波也相当风光。而德米斯·哈萨比斯才刚在X上祝贺了自己的前同事杰弗里·辛顿，结果一天后自己也获奖了），这被大家笑称今年诺贝尔奖是狠狠蹭了一下AI的热度。
但相比物理学奖给得还略有些理由牵强，化学奖倒是争议不大，三位科学家的确极大程度改变了蛋白质的设计和对蛋白质结构的获取流程，并且他们还通过模型开源，直接推动了生物学、药物开发等相关研究领域可以极大地缩短工作周期，这对人类带来的贡献无法估量。
一定要说AI的话，是AI的确已经相当程度地介入了我们的生产生活，并且它完全可以在所有领域都发挥作用，两位DeepMind的科学家以AI的方式解决了化学方面的问题，其他科学家也可以用同样的方式解决其它领域的其它问题。
正所谓“在AI时代，很多产品都可以重做一次”，或许同样的，在AI时代，很多课题也都可以再研究一次。接下来我们看到更多“被AI加持”的科学家获得更多的诺贝尔奖也未可知。
本文图片来自慕尼黑工业大学、美国国家科学院院刊、诺贝尔奖官方。 #今年诺贝尔化学奖授予了蛋白质设计和结构预测而这也是 #的再一次胜利
英文版:
On October 9, 2024, at 5:45 p.m. Beijing time, the 2024 Nobel Prize in Chemistry was announced. The Royal Swedish Academy of Sciences awarded this year's prize to David Baker, Demis Hassabis, and John Michael Jumper for their research on protein structure and design.
The three scientists will share a prize of 11 million Swedish kronor. Half of the prize is awarded to David Baker for his contribution to computational protein design, while the other half is shared equally by Demis Hassabis and John Michael Jumper for their contribution to protein structure prediction.
David Baker, a biochemist and computational biologist, was born on October 6, 1962, in Seattle, Washington. He received his Ph.D. in biochemistry from the University of California, Berkeley in 1989 and is currently a fellow of the National Academy of Sciences and the director of the Institute for Protein Design at the University of Washington.
Prior to the 1990s, scientists could only modify existing proteins to give them new functionalities, such as acting as catalysts in the chemical industry. David Baker and his team chose to design proteins from scratch by drawing entirely new structures. They used their proprietary software, Rosetta, to calculate the necessary amino acid sequences and achieved success.
This breakthrough marked a significant development in the field of protein engineering. The ability to directly design proteins through computational methods and algorithms, without being limited by existing natural proteins, empowered scientists to custom-design proteins for applications in disease treatment, drug development, biology, materials science, and more.
Subsequently, David Baker open-sourced Rosetta, further propelling global scientists to continuously apply and iterate on Rosetta, directly influencing the work of Demis Hassabis and John Michael Jumper.
Demis Hassabis, born on July 27, 1976, in London, earned a Ph.D. in cognitive neuroscience from University College London. He conducted research in neuroscience and artificial intelligence at the Massachusetts Institute of Technology and Harvard University. In 2010, he co-founded DeepMind with Shane Legg and Mustafa Suleyman. After DeepMind was acquired by Google in 2014, he became CEO of DeepMind.
John Michael Jumper, born in 1985 in Little Rock, Arkansas, earned his Ph.D. from the University of Chicago in 2017. He currently serves as the director and senior research scientist at DeepMind. It is quite remarkable that he received the Nobel Prize at such a young age.
Further reading: In 1915, William Lawrence Bragg won the Nobel Prize in Physics at the age of 25, making him the youngest recipient of a Nobel Prize in the natural sciences.
These two individuals are not traditional scientists in the fields of chemistry or biology. DeepMind’s focus on protein structure prediction was primarily to validate the capability of artificial intelligence. Following the success of AlphaGo (a project by DeepMind that defeated Lee Sedol and Ke Jie in the game of Go), protein structure prediction became the next research topic.
After Demis Hassabis introduced AlphaFold, the accuracy of protein structure prediction significantly improved. John Michael Jumper joined the team and together with Demis Hassabis, led the development of AlphaFold2. This deep learning algorithm can predict protein structures with remarkable accuracy, revolutionizing the conventional approach in the scientific community over the past few decades. This has become the new standard.
AlphaFold2 was later open-sourced by DeepMind and has been widely adopted by researchers globally in their daily studies (note that such predictions are not entirely accurate but help researchers assess the reliability of results. Traditionally, determining the structure of a protein could take several years or may not even be successful, but with AlphaFold2, it can be completed in just a few minutes).
Interestingly, following the awarding of the Nobel Prize in Physics to two pioneers in the field of machine learning the day before, this year's Chemistry Prize was also granted to two big names in the AI field (which showcased Google's prominence in this area). Demis Hassabis had just congratulated his former colleague, Geoffrey Hinton, on his award at X, only to receive his own award a day later. This led to jokes that this year's Nobel Prize awards capitalized on the popularity of AI.
However, unlike the Physics Prize, which had some controversy, the Chemistry Prize was relatively uncontroversial. The significant impact of these three scientists on protein design and structure acquisition processes cannot be underestimated. By open-sourcing their models, they directly accelerated research in biology, drug development, and related fields, greatly reducing work cycles and making invaluable contributions to humanity.
It is undeniable that AI has significantly integrated into our production and daily lives and has the potential to be functional across all fields. These two DeepMind scientists solved a chemistry problem using AI, suggesting that other scientists can also address problems in different fields using similar methods.
"In the era of AI, many products can be redesigned," and perhaps similarly, in the AI era, many subjects can be revisited for further study. It remains uncertain whether we will witness more Nobel laureates supported by AI in the future.
Image credits: Technical University of Munich, Proceedings of the National Academy of Sciences, and the Nobel Prize official website.
今年诺贝尔化学奖授予了蛋白质设计和结构预测，而这也是 AI 的再一次胜利
#今年诺贝尔化学奖授予了蛋白质设计和结构预测而这也是 #的再一次胜利
关注流程：打开随信App→搜索新湖畔网随信号：973641 →订阅即可！
公众号：新湖畔网 抖音：新湖畔网
视频号：新湖畔网 快手：新湖畔网
小红书：新湖畔网 随信：新湖畔网
百家号：新湖畔网 B站：新湖畔网
知乎：新湖畔网 微博：新湖畔网
UC头条:新湖畔网 搜狐号：新湖畔网
趣头条：新湖畔网 虎嗅：新湖畔网
腾讯新闻：新湖畔网 网易号：新湖畔网
36氪：新湖畔网 钛媒体：新湖畔网
今日头条：新湖畔网 西瓜视频：新湖畔网
