Machine learning applications for therapeutic tasks with genomics data

Area of Science:

• Biomedical Informatics
• Computational Biology
• Genomics

Background:

• Increasing availability of genomics and biomedical data fuels machine learning (ML) applications in drug discovery.
• ML algorithms offer powerful tools for analyzing complex biological datasets.
• Therapeutic development relies on integrating diverse data types for effective drug design and clinical application.

Purpose of the Study:

• To survey machine learning applications in genomics within the context of therapeutic development.
• To explore the integration of genomics with other data modalities like electronic health records and clinical texts.
• To identify current challenges and future research directions in ML for genomics-driven therapeutics.

Main Methods:

• Systematic literature review of machine learning applications in genomics for drug discovery and development.
• Analysis of the interplay between genomics, compounds, proteins, electronic health records, cellular images, and clinical texts.
• Categorization of ML applications across the entire therapeutics pipeline.

Main Results:

• Identified 22 distinct machine learning in genomics applications spanning the full therapeutics pipeline.
• Investigated the integration of genomics with diverse data sources including EHRs, imaging, and text data.
• Highlighted key challenges and areas for future expansion in the field.

Conclusions:

• Machine learning significantly enhances various stages of therapeutic development, from target identification to post-market analysis.
• The integration of multi-modal data with genomics presents a promising avenue for advancing precision medicine.
• Addressing identified challenges is crucial for unlocking the full potential of ML in genomics for future therapeutics.