Enhancing mRNA translation efficiency with discriminative and generative artificial intelligence by optimizing 5' UTR...

Area of Science:

• Biotechnology
• Computational Biology
• Molecular Biology

Background:

• Messenger RNA (mRNA) therapeutics, including mRNA vaccines, offer powerful disease-fighting capabilities.
• A key limitation for exogenous mRNA application is its inherently low translation efficiency.
• Optimizing mRNA sequences is crucial for advancing therapeutic development.

Purpose of the Study:

• To develop a computational framework, UTailoR, for optimizing 5' UTR sequences to enhance mRNA translation efficiency.
• To improve the effectiveness and applicability of mRNA-based therapeutics.
• To provide an accessible online tool for mRNA 5' UTR sequence optimization.

Main Methods:

• A two-step artificial intelligence strategy was employed, integrating deep learning and generative modeling.
• A deep-learning-based discriminative model was developed to predict mRNA translation efficiency based on 5' UTR sequences.
• A generative model was utilized to produce optimized 5' UTR sequences closely resembling original sequences but with enhanced translation efficiency.

Main Results:

• The UTailoR framework successfully optimized 5' UTR sequences, leading to significantly improved mRNA translation efficiency.
• Optimized sequences demonstrated approximately a 200% increase in translation efficiency compared to original sequences.
• The computational approach proved efficient and convenient for mRNA 5' UTR optimization.

Conclusions:

• UTailoR offers an effective computational solution for enhancing mRNA translation efficiency through 5' UTR sequence optimization.
• This advancement holds significant potential for improving the performance of mRNA therapeutics and vaccines.
• The availability of UTailoR online facilitates wider adoption and application in the field of mRNA technology.