Clinical Variant Databases and Machine Learning Prediction Supporting Genomic Medicine.

Area of Science:

• Genomics
• Bioinformatics
• Machine Learning

Background:

• Genomic medicine, utilizing genomic variations for clinical decisions, is increasingly accessible.
• Accurate interpretation of genomic variations is vital for effective genomic medicine.
• Current genomic data lacks diversity, biasing interpretation and impacting certain ethnic groups.

Purpose of the Study:

• To outline the workflow of genomic medicine and variant interpretation.
• To introduce disease variant databases for advancing genomic medicine.
• To highlight the need for improved machine learning methods for noncoding variant prediction.

Main Methods:

• Review of genomic medicine processes.
• Introduction of established disease variant databases.
• Discussion of machine learning approaches for variant effect prediction.

Main Results:

• Genomic data biases necessitate enhanced diversity for equitable healthcare.
• Machine learning models for coding variants are advanced, but noncoding variant prediction requires improvement.
• Accessible genomic medicine relies on robust variant interpretation tools.

Conclusions:

• Improving genomic data diversity and noncoding variant prediction is crucial for equitable and effective genomic medicine.
• Advanced variant interpretation, including machine learning, is key to realizing the potential of genomic medicine.
• This chapter provides a foundation for understanding genomic medicine workflows and future research directions.