I-Corps: Developing predictive computational methods for drug development

这个I-Corps项目的更广泛的影响/商业潜力是为制药和生物技术行业提供一个软件或服务，为整个生命科学的实验产生准确的预测。药物发现、生物制剂设计和其他工业生物技术努力往往需要数十亿美元和数十年的研究才能完成项目。尽管在过去的二十年中，使用计算工具加速药物发现过程一直是一个主要的焦点，但现有技术的功效通常仅限于具有大量精心策划的数据的少数实验。提出的技术可以为临床前发现空间的实验提供准确的预测，解决了其他当代技术范围之外的几个步骤。该技术的灵活性允许计算机执行几个时间和资源密集的步骤，以增加实验吞吐量，同时使研究人员能够专注于高成功率的实验。这种改善时间和资源管理的机会可能会加快药物进入临床的时间，并可能激发对治疗不足疾病的治疗努力。这个I-Corps项目将侧重于确定当前计算方法在制药和生物技术领域的哪些实验服务不足，从而解决药物发现中的这些瓶颈。这项技术利用了一种利用生物物理模拟信息生成人工智能（AI）模型的新方法。这种人工智能方法大大减少了生成准确预测所需的数据量，与传统人工智能方法相比，具有更大的实用范围。通过几项回顾性研究，证明了该技术能够为小分子-蛋白质相互作用、蛋白质-蛋白质相互作用、治疗肽稳定等提供准确预测。该技术的进一步探索需要从制药和生物技术领域的潜在客户那里获得洞察力，以了解该技术需要解决的具体实验，从而为未来的前瞻性研究提供信息。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this I-Corps project is to provide the pharmaceutical and biotechnology industries with a software or service that generates accurate predictions for experiments across the life sciences. Drug discovery, biologics design, and other industrial biotechnological efforts often require billions of dollars and decades of research to see projects to completion. Although there has been a major focus over the past two decades to accelerate the drug discovery process using computational tools, the efficacy of existing technologies has typically been limited to a select few experiments which have large amounts of curated data. The proposed technology can provide accurate predictions for experiments across the preclinical discovery space, addressing several steps outside the purview of other contemporary technologies. The flexibility of this technology allows for several time and resource intensive steps to be performed by computer to increase experimental throughput, while allowing researchers to focus on experiments with a high probability for success. This opportunity for improved time and resource management may accelerate the timeline for drugs to enter the clinic and may invigorate therapeutic efforts towards underserved diseases.This I-Corps project will focus on identifying which experiments within the pharmaceutical and biotechnology space are underserved by current computational methods, allowing for these bottlenecks within drug discovery to be addressed. This technology makes use of a novel approach for generating artificial intelligence (AI) models employing information from biophysical simulations. This AI method significantly reduces the amount of data required for generating accurate predictions and has an increased scope of utility compared with traditional AI methods. Proof of concept has been demonstrated through several retrospective studies which demonstrated the technology's ability to provide accurate predictions toward small molecule-protein interactions, protein-protein interactions, therapeutic peptide stabilization, and more. Further exploration of the technology requires insight from potential customers in the pharmaceutical and biotechnology space to understand the specific experiments that this technology needs to address to inform future prospective studies.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.