Zheng Jiayao, Wang Yu, Liang Qianying, Cui Lun, Wang Liqun
School of Pharmacy & School of Biological and Food Engineering, Changzhou University, Changzhou 213164, China.
Protein Design Lab, Changzhou AiRiBio Healthcare Co., Ltd., Changzhou 213164, China.
Molecules. 2024 Dec 16;29(24):5923. doi: 10.3390/molecules29245923.
Antibodies play critical roles in modern medicine, serving as diagnostics and therapeutics for various diseases due to their ability to specifically bind to target antigens. Traditional antibody discovery and optimization methods are time-consuming and resource-intensive, though they have successfully generated antibodies for diagnosing and treating diseases. The advancements in protein data, computational hardware, and machine learning (ML) models have the opportunity to disrupt antibody discovery and optimization research. Machine learning models have demonstrated their abilities in antibody design. These machine learning models enable rapid in silico design of antibody candidates within a few days, achieving approximately a 60% reduction in time and a 50% reduction in cost compared to traditional methods. This review focuses on the latest machine learning-based antibody discovery and optimization developments. We briefly discuss the limitations of traditional methods and then explore the machine learning-based antibody discovery and optimization methodologies. We also focus on future research directions, including developing Antibody Design AI Agents and data foundries, alongside the ethical and regulatory considerations essential for successfully adopting machine learning-driven antibody designs.
抗体在现代医学中发挥着关键作用，由于其能够特异性结合靶抗原，因而可作为多种疾病的诊断和治疗手段。传统的抗体发现和优化方法耗时且资源密集，尽管它们已成功产生用于疾病诊断和治疗的抗体。蛋白质数据、计算硬件和机器学习（ML）模型的进步有机会颠覆抗体发现和优化研究。机器学习模型已在抗体设计中展示了其能力。这些机器学习模型能够在几天内快速进行抗体候选物的计算机模拟设计，与传统方法相比，时间减少了约60%，成本降低了50%。本综述重点关注基于机器学习的抗体发现和优化的最新进展。我们简要讨论传统方法的局限性，然后探索基于机器学习的抗体发现和优化方法。我们还关注未来的研究方向，包括开发抗体设计人工智能代理和数据铸造厂，以及成功采用机器学习驱动的抗体设计所必需的伦理和监管考量。
