ASOptimizer: optimizing chemical diversity of antisense oligonucleotides through deep learning

Area of Science:

• Biochemistry
• Bioinformatics
• Genetics

Background:

• Antisense oligonucleotides (ASOs) offer a promising avenue for gene expression modulation in therapeutic applications.
• Manual design of ASOs is a complex, time-consuming, and resource-intensive process, hindering rapid research and development.

Purpose of the Study:

• To introduce ASOptimizer, a user-friendly web server utilizing a deep learning framework to optimize ASO sequences and chemical modifications.
• To provide researchers with an accessible computational tool for efficient ASO design, reducing the need for specialized expertise.

Main Methods:

• Development of a deep learning-based computational framework, ASOptimizer.
• Implementation of a web server interface for ASOptimizer, allowing users to input ASO sequences.
• Systematic exploration of chemical modification sites within nucleic acid sequences.

Main Results:

• The ASOptimizer web server systematically explores modification sites and provides a ranked list of optimized ASO sequences and modification patterns.
• The platform offers an intuitive interface, making advanced ASO optimization accessible without requiring deep learning expertise.

Conclusions:

• ASOptimizer significantly simplifies and accelerates the design of effective antisense oligonucleotides.
• The web server democratizes access to sophisticated ASO optimization tools, fostering broader advancements in gene therapy research.