He Won the Nobel Prize for Protein Design. Now He Uses AI to Create Molecules Not Found in Nature
He Won the Nobel Prize for Protein Design. Now He Uses AI to Create Molecules Not Found in Nature
他因蛋白质设计获得诺贝尔奖，如今正利用人工智能创造自然界中不存在的分子
The natural world as we know it represents only a fraction of what might exist. Based on this idea, AI BioDesign was born: a scientific project that combines artificial intelligence and large-scale laboratory experiments to design and test new molecules and biological functions that are not found in nature but are physically and chemically possible. 我们所知的自然界仅代表了可能存在的事物的一小部分。基于这一理念，“AI BioDesign”项目应运而生：这是一个结合了人工智能与大规模实验室实验的科学项目，旨在设计并测试那些自然界中不存在，但在物理和化学上却完全可行的全新分子和生物功能。
In doing so, the researchers are aiming to create databases, models, and tools that could serve as “seeds” for developing the medicines and technologies of the future. The authors envision, for example, new drugs to treat cancer and neurodegenerative diseases or enzymes capable of breaking down plastics in the ocean. 通过这项工作，研究人员旨在建立数据库、模型和工具，作为开发未来药物和技术的“种子”。例如，研究人员设想开发出治疗癌症和神经退行性疾病的新药，或是能够分解海洋塑料的酶。
Proteins that do not exist in nature but are technically possible can be designed and constructed using AI. ILLUSTRATION: Ian C. Haydon / UW Medicine Institute for Protein Design 利用人工智能可以设计并构建出自然界中不存在但在技术上可行的蛋白质。图片来源：Ian C. Haydon / 华盛顿大学医学院蛋白质设计研究所
The initiative is led by the Allen Institute, a biomedical research nonprofit in Seattle, along with the University of Washington and the Fred Hutchinson Cancer Center. Among the scientists leading the charge is David Baker, who shared the 2024 Nobel Prize in Chemistry for his work in computational protein design. 该计划由位于西雅图的生物医学研究非营利组织艾伦研究所（Allen Institute）、华盛顿大学以及弗雷德·哈钦森癌症中心共同领导。领衔该项目的科学家之一是大卫·贝克（David Baker），他因在计算蛋白质设计方面的贡献而共同获得了2024年诺贝尔化学奖。
For much of the history of biology, scientists have studied the solutions forged by billions of years of evolution. Exploring possibilities that nature never produced is a fundamentally new approach. And as technology makes this now possible, one that a leaders of the field even compares the potential of these methods to historic transformations such as industrialization, electrification, and the digital revolution. 在生物学历史的大部分时间里，科学家们研究的都是经过数十亿年进化所形成的解决方案。探索自然界从未产生过的可能性是一种根本性的新方法。随着技术使之成为可能，该领域的一些领军人物甚至将这些方法的潜力与工业化、电气化和数字革命等历史性变革相提并论。
However, delving into completely novel regions of biology also inevitably leads to difficult questions. WIRED en Español spoke with David Baker, chief scientific officer at AI BioDesign, to address this and other topics. This interview has been edited for length and clarity. 然而，深入探索生物学中全新的领域也必然会带来棘手的问题。《连线》西班牙语版（WIRED en Español）采访了 AI BioDesign 的首席科学官大卫·贝克，探讨了这一话题及其他相关议题。为篇幅和清晰起见，本次采访内容经过了编辑。
JORGE GARAY: AI BioDesign seeks to explore possible molecules and biological functions that have never existed in nature. Are there any particular risks involved in venturing into this uncharted territory? How can scientists anticipate those risks and what steps can they take to minimize them before a designed molecule leaves the lab? 豪尔赫·加雷（JORGE GARAY）：AI BioDesign 旨在探索自然界中从未存在过的潜在分子和生物功能。进入这一未知领域是否存在特殊的风险？科学家们如何预见这些风险，并在设计的分子离开实验室之前采取什么措施来将其降至最低？
DAVID BAKER: What’s exciting about biology is that nature has explored only a fraction of what is physically and chemically possible, which opens a world of possibilities for what we could design. When we explore those possibilities, the primary risks are not much different than those associated with any new biological technology—unintended interactions with living systems, unexpected environmental effects, or misuse. 大卫·贝克： 生物学最令人兴奋之处在于，自然界仅探索了物理和化学上可能存在的事物的一小部分，这为我们的设计开启了一个充满可能性的世界。当我们探索这些可能性时，主要的风险与任何新技术所带来的风险并无太大不同——即与生命系统的意外相互作用、意想不到的环境影响或被滥用。
The advantage we have today is that computational design allows us to evaluate many of these risks before a molecule is ever synthesized. We can screen designs computationally, test them extensively in contained laboratory settings, and subject them to increasingly realistic experimental validation before considering any real-world application. 我们今天的优势在于，计算设计使我们能够在分子合成之前就评估其中的许多风险。我们可以在计算机上筛选设计，在封闭的实验室环境中进行广泛测试，并在考虑任何实际应用之前，对其进行日益逼真的实验验证。
David Baker, winner of the 2024 Nobel Prize in Chemistry. Courtesy of the Allen Institute 2024年诺贝尔化学奖得主大卫·贝克。图片由艾伦研究所提供
The project mentions possibilities ranging from new drugs to plastic-degrading enzymes and even biological computers. If we have greater predictive power when designing new molecules and biological functions, how far do you think this capability could take us? 该项目提到了从新药到塑料降解酶，甚至生物计算机等多种可能性。如果我们能在设计新分子和生物功能时拥有更强的预测能力，你认为这种能力能将我们带向何方？
My team and I have lofty goals for protein design: We’re working to build a world where a cure for a new disease is created in weeks, not decades. Where our air, water, and soil are clean because we removed pollutants and reimagined the processes that contaminated them. Where crops thrive in conditions that once killed them. Where molecular machines pull critical minerals from waste and repair the infrastructure we depend on. 我和我的团队对蛋白质设计有着宏伟的目标：我们致力于构建一个新疾病的治疗方法能在几周而非几十年内被研发出来的世界。在这个世界里，我们的空气、水和土壤因去除了污染物并重塑了污染过程而变得洁净；农作物能在曾经导致它们死亡的环境中茁壮成长；分子机器能从废弃物中提取关键矿物质，并修复我们所依赖的基础设施。
Proteins are the molecular machines life has developed to create all organic matter we know of here on Earth. Unlocking their full potential requires that we derive engineering principles that make it possible for innovative new ideas, including ones we can’t yet imagine, to be achievable. 蛋白质是生命进化出的分子机器，创造了我们在地球上所知的所有有机物质。要释放它们的全部潜力，我们需要推导出工程原理，使创新的新想法——包括我们目前还无法想象的想法——成为可能。
As these models become more capable, could AI eventually design functional molecules that scientists themselves do not fully understand? If so, would it be enough to experimentally demonstrate that they work and are safe, or do you think we also need to understand the mechanisms? 随着这些模型能力越来越强，人工智能最终是否会设计出科学家自己都无法完全理解的功能分子？如果是这样，仅通过实验证明它们有效且安全是否足够，还是说你认为我们还需要理解其背后的机制？
Science has often progressed in stages, and we see that mirrored in how machine learning has advanced protein design. The first step is usually observing that something works, and we often only later understand why. Machine learning is really good at that first stage—it has vastly improved our capability to observe patterns that can be leveraged for biological design. Strong experimental evidence can justify moving forward with projects, but deeper understanding remains an important objective both for our research and scientific inquiry overall. Much of what we cannot achieve today is due to a lack of understanding or data that defines the parameters for how specific systems work or do not. This is the crux of the AI BioDesign program. 科学往往是分阶段进步的，我们在机器学习推动蛋白质设计的方式中看到了这一点。第一步通常是观察到某件事物有效，而我们往往在事后才理解其原因。机器学习非常擅长第一阶段——它极大地提高了我们观察可用于生物设计模式的能力。强有力的实验证据可以证明项目推进的合理性，但更深入的理解仍然是我们研究和整体科学探索的重要目标。我们今天无法实现的许多事情，都是由于缺乏理解或缺乏定义特定系统如何运作或为何不运作的参数数据。这正是 AI BioDesign 项目的核心所在。
All proteins are built of the same amino acids, but they can be combined in innumerable ways to create never-before-seen structures. ILLUSTRATION: Sanjay Srivatsan/Fred Hutchinson Cancer Center 所有蛋白质都由相同的氨基酸构成，但它们可以通过无数种方式组合，创造出前所未有的结构。图片来源：Sanjay Srivatsan / 弗雷德·哈钦森癌症中心
Are there molecules or biological functions that, in your opinion, should not be designed, even if it were technically possible to do so? What criteria should determine where that line is drawn? 在你看来，是否有些分子或生物功能即使在技术上可行，也不应该被设计出来？应该用什么标准来划定这条界限？
Decisions should be guided by a balance of potential benefits and potential harms. Applications that address major challenges in health, sustainability, or human well-being have a strong case for development. Conversely, designs that create significant risks to public safety, security, or the environment deserve heightened scrutiny and, in some cases, clear restrictions. 决策应以潜在利益与潜在危害之间的平衡为指导。那些能够解决健康、可持续性或人类福祉方面重大挑战的应用，具有充分的开发理由。相反，那些对公共安全、国家安全或环境造成重大风险的设计，则应受到更严格的审查，在某些情况下，还应受到明确的限制。
Just as we have developed frameworks for other powerful technologies, we need governance structures that evolve alongside advances in AI and biotechnology. 正如我们为其他强大技术制定了框架一样，我们需要与人工智能和生物技术的进步同步发展的治理结构。