A novel deep generative model for mRNA vaccine development: Designing 5′ UTRs with N1-methyl-pseudouridine modification
Xiaoshan Tang, Miaozhe Huo, Yuting Chen, Hai Huang, Shugang Qin, Jia‐Qi Luo, Zeyi Qin, Xin Jiang, Yongmei Liu, Xing Duan, Ruohan Wang, Lingxi Chen, Hao Li, Na Fan, Zhongshan He, Xi He, Bairong Shen, Shuai Cheng Li, Xiangrong Song
Abstract
Efficient translation mediated by the 5ʹ untranslated region (5ʹ UTR) is essential for the robust efficacy of mRNA vaccines. However, the N1-methyl-pseudouridine (m1Ψ) modification of mRNA can impact the translation efficiency of the 5ʹ UTR. We discovered that the optimal 5ʹ UTR for m1Ψ-modified mRNA (m1Ψ–5ʹ UTR) differs significantly from its unmodified counterpart, highlighting the need for a specialized tool for designing m1Ψ–5ʹ UTRs rather than directly utilizing high-expression endogenous gene 5ʹ UTRs. In response, we developed a novel machine learning-based tool, Smart5UTR, which employs a deep generative model to identify superior m1Ψ–5ʹ UTRs in silico. The tailored loss function and network architecture enable Smart5UTR to overcome limitations inherent in existing models. As a result, Smart5UTR can successfully design superior 5ʹ UTRs, greatly benefiting mRNA vaccine development. Notably, Smart5UTR-designed superior 5ʹ UTRs significantly enhanced antibody titers induced by COVID-19 mRNA vaccines against the Delta and Omicron variants of SARS-CoV-2, surpassing the performance of vaccines using high-expression endogenous gene 5ʹ UTRs.