NSF Center for Computer-Assisted Synthesis

NSF计算机辅助合成中心(CCAS)是合作、创新和教育的纽带，将数据科学和化学合成结合在一起。由合成有机化学家、计算化学家和计算机科学家组成的高度跨学科的CCAS团队正在开发数据科学工具和计算工作流，这些工具和计算工作流可能会塑造合成化学及其实现的领域的未来，如医学、材料科学和能源研究。由学术、工业和非营利性合作伙伴和研究中心组成的广泛网络进一步放大了该网站的影响，其数据化学工具正通过开源信息交换所与研究界共享。所有这些都为CCAS提供了一个独特的机会来开发、交流和评估数据化学领域的想法，其共享的工具和培训将使学生、实习化学家和化学工业能够有效地将数据科学应用于他们自己的化学研究。CCAS由每个阶段的有机化学家授课，专注于受使用启发的数据科学研究，推动新的数据类型和机器学习(ML)方法的发展，使之能够发现新的反应并产生新的科学见解。这四项科学努力包括(I)开发有效的ML工具来优化化学反应，(Ii)通过可解释的统计模型和电子结构计算获得机理上的理解，(Iii)预测反应结果以预测和发现新的反应性，以及(Iv)整合这些工具以有效地计划和执行复杂分子的多步合成。为了实现这些目标，三个主题交织在每个推进中：(A)新的结构化数据类型，这些数据类型能够从头开始接受高通量实验和预测模型，超越了常用数据库中的信息；(B)连接基于描述符和基于结构的深度学习范式的分子和反应表示；以及(C)专门为整个化学中普遍存在的低数据制度设计的算法。通过这些综合的研究主题和推进，CCAS构建和共享数据化学平台，预计这些平台将使化学家能够应对该领域目前装备不足的雄心勃勃的挑战。数据化学平台还将在本科生和研究生教育中开辟新的机会，通过与数据化学家网络的合作伙伴关系和残疾化学家的研究机会，CCAS寻求扩大来自代表性不足群体的研究人员的参与。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。

The NSF Center for Computer Assisted Synthesis (CCAS) is a nexus of collaboration, innovation, and education that brings together data science and chemical synthesis. The highly interdisciplinary CCAS team, composed of synthetic organic chemists, computational chemists, and computer scientists, is developing data science tools and computational workflows that will likely shape the future of synthetic chemistry and the fields it enables, such as medicine, materials science, and energy research. This site’s impacts are being further amplified by an extensive network of academic, industrial and non-profit partners and research centers, and its data chemistry tools are being shared with the research community through open-source clearinghouses. All of this provides CCAS with a unique opportunity to develop, exchange, and evaluate ideas in the field of data chemistry, and its shared tools and training will empower students, practicing chemists, and the chemical industry to effectively apply data science to their own chemical research.Led by organic chemists at every stage, CCAS focuses on use-inspired data science research that drives the development of new data types and machine learning (ML) methods that enable the discovery of novel reactions and yield new scientific insights. The four scientific thrusts include (i) developing effective ML tools for optimizing chemical reactions, (ii) gaining mechanistic understanding through interpretable statistical models and electronic structure calculations, (iii) predicting reaction outcomes to anticipate and discover new reactivity and (iv) integrating these tools for the efficient planning and execution of multistep syntheses of complex molecules. To accomplish these goals, three themes are interwoven into each of the thrusts: (a) new structured data types that are amenable to high-throughput experimentation and predictive models from the ground up, going beyond the information from commonly used databases, (b) molecular and reaction representations that bridge descriptor-based and structure-based deep learning paradigms, and (c) algorithms specifically designed for the low data regimes prevalent throughout chemistry. Through these integrated research themes and thrusts, CCAS constructs and shares data chemistry platforms that are expected to enable chemists to tackle ambitious challenges that the field is currently under-equipped to pursue. The data chemistry platform also will open up new opportunities in undergraduate and graduate education, and through partnerships with the Data Chemists Network and research opportunities for chemists with disabilities, CCAS seeks to broaden the participation of researchers from underrepresented groups.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.