CAREER: Error-Free, Uniform and Composable Chemical Computation

通过将以前需要人工干预的任务自动化，电子计算机的计算已经彻底改变了生活中相对平凡的方面，比如购物和通勤。同样，化学编程--自动处理嵌入在分子中的信息--可能会重塑世界：由可编程化学反应控制的智能分子可以在分子水平上实现对物质结构的精确自动控制。例如，有朝一日，这可能会给应用于活细胞或自组装纳米材料内的治疗方法带来革命性的变化。该项目旨在发展这种化学信息处理的数学基础，使基于化学的“软件”工程更接近现代电子计算的可靠性。研究将针对该领域的实践者所面临的关键挑战：即防错、可重用性以及将单独设计的化学系统整合为一个正常运行的整体的能力。从经典计算机科学的研究中收集的见解将用于实现这些目标，但需要基于物理和化学定律的新见解来推理调解化学计算的独特的分子相互作用。在医学或纳米工程中的潜在应用包括识别和治疗疾病，以及纳米级精度的自组装设备。这个项目将帮助培养新一代跨学科的科学家和程序员，他们能够创新出强大的纳米级信息技术。这个项目为编程化学算法--即人工合成化学反应执行的算法--提供了理论基础，确保它们具有现有系统所缺乏的，或者充其量是知之甚少的三个关键特性：(1)无错误：可由忠实地执行预期算法的真实化学物质实现，(2)均匀：对任何“群体大小”，即分子总数都是正确的，不像许多目前的算法，反应必须专门针对群体大小量身定做；(3)可合成：可以封装成易于组合的功能模块。随着DNA纳米技术和分子计算的成熟，我们也必须洞察它们的基本能力和局限性。严谨的化学计算理论将指导未来的实验突破。这一理论的发展将以对底物独立的偏好为指导，确定所有化学系统遵守的计算定律，从它们共同的统一原则中筛选出特定技术的人工制品。使化学算法无错误、统一和可组合的技术的发展可能会导致一条编制化学程序的系统路径，使其可被非化学家理解和使用，从而使分子计算革命得以发展。在实现自动催化反应的可编程动力学障碍方面的进展最终可能导致重大技术突破，从而实现错误恢复的单分子检测。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Computation by electronic computers has revolutionized relatively mundane aspects of life, such as shopping and commuting, by automating tasks that previously required human intervention. Similarly, the programming of chemicals-automating the processing of information embedded in molecules could remake the world: smart molecules controlled by programmable chemical reactions could achieve the same level of precise automated control over the configuration of matter at the molecular level. This may one day revolutionize, for example, therapeutic treatments applied within living cells or nanoscale materials fabricated by self-assembly. This project aims to develop the mathematical foundations of such chemical information processing, bringing the engineering of chemistry-based "software" closer to the reliability of modern electronic computing. Research will be directed toward key challenges that have faced practitioners in the field: namely error-prevention, reusability, and the ability to integrate separately designed chemical systems into a properly functioning whole. Insights gleaned from research in classical computer science will be applied to achieve these goals, but new insights, based on the laws of physics and chemistry, will be required to reason about the uniquely molecular interactions mediating chemical computation. The potential applications in medicine or nano-engineering include identifying and curing diseases, as well as self-assembling devices with nanoscale precision. Together with a tightly integrated educational plan based on training female undergraduate mathematics students in computer science research, and a data-driven approach to develop autograding software for undergraduate CS courses, this project will help train a new and diverse generation of interdisciplinary scientists and programmers, who can innovate robust nanoscale information technologies.This project advances the theoretical foundation for programming chemical algorithms-that is, algorithms executed by artificially synthesized chemical reactions-ensuring they have three crucial properties that are lacking, or at best poorly understood, in existing systems: (1) error-free: implementable by real chemicals that faithfully execute the intended algorithm, (2) uniform: correct for any "population size", i.e., the total number of molecules, unlike many current algorithms where reactions must be tailored specifically to the population size and (3) composable: can be packaged into functional modules that are easily combined. As DNA nanotechnology and molecular computing mature, so too must our insight into their fundamental abilities and limitations. A rigorous theory of chemical computing will guide the experimental breakthroughs of the future. Development of this theory will be guided by a preference for substrate-independence, identifying the laws of computation obeyed by all chemical systems, sifting out artifacts of particular technologies from their shared unifying principles. The development of techniques to make chemical algorithms error-free, uniform, and composable could lead to a systematic path for programming chemistry, making it understandable and usable by non-chemists thus allowing the molecular computing revolution to take flight. Advances in programmable kinetic barriers in implementing an autocatalytic reaction could eventually lead to a significant technological breakthrough allowing error-resilient single-molecule detection.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.