AI-based platform for predicting emerging vaccine-escape variants and designing mutation-proof antibodies

项目摘要 由于大规模接种疫苗，2019年冠状病毒病（COVID-19）大流行引起的 严重急性呼吸道综合征冠状病毒2型（SARS-CoV-2）已得到部分控制。 然而，新出现的传染性变异，如德尔塔，仍在助长新一波的感染 世界各地疫苗逃逸（或疫苗突破）变异对人类构成新的威胁。 我们与COVID-19的斗争了解病毒的突变和进化是非常重要的 重要性通过整合基因组分析，人工智能（AI），计算生物物理学， 高等数学，实验数据，PI已经建立了一个全面的计划， 预测SARS-CoV-2的实验水平准确度和群体水平可靠度 变异感染性和抗体破坏。预测未来新兴市场仍然具有挑战性， 疫苗逃逸变体，开发下一代疫苗，设计突变- 证明抗体疗法。这些挑战在拟议的项目中得到解决。新 将开发数学工具和人工智能算法，以进一步改善目前的状况， 最先进的预测突变引起的病毒感染性变化，疫苗突破， 抗体破坏。未来新出现的变异中的重要突变将根据以下因素进行预测： 分子机制、自然选择和进化效应。新型抗突变抗体 药物的设计和测试将基于这些抗体，这些抗体已经通过了早期的临床 审判这些预测模型将在一个用户友好的平台上实施，并配备在线服务器 让研究人员设计出抗突变的新疫苗和抗体疗法。拟议 方法将被应用于预测流感中出现的变种，并提高 季节性流感疫苗

Project Summary Due to massive vaccination, coronavirus disease 2019 (COVID-19) pandemic caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been partially under control. However, emerging contagious variants such as Delta are still fueling new waves of infections around the world. Vaccine-escape (or vaccine-breakthrough) variants pose renewed threats to our battle against COVID-19. Understanding viral mutagenesis and evolution is of preeminent importance. By integrating genomic analysis, artificial intelligence (AI), computational biophysics, advanced mathematics, and experimental data, the PIs have built a comprehensive program with the experimental level of accuracy and population-level of reliability for predicting SARS-CoV-2 variant infectivity and antibody disruption. It remains challenging to forecast future emerging vaccine-escape variants, to develop the next-generation of vaccines, and to design mutation- proof antibody therapeutics. These challenges are tackled in the proposed project. New mathematical tools and AI algorithms will be developed to further improve the current state-of- the-art in predicting mutation-induced viral infectivity changes, vaccine breakthroughs, and antibody disruptions. Vital mutations in future emerging variants will be forecasted based on molecular mechanisms, natural selection, and evolutionary effects. New mutation-proof antibody drugs will be designed and tested based on those antibodies that had gone through earlier clinical trials. The predictive models will be implemented into a user-friendly platform with online servers for researchers to design mutation-proof new vaccines and antibody therapies. The proposed methods will be applied to forecast emerging variants in the flu and improve the efficacy of seasonal flu vaccines.