RExPRT: a machine learning tool to predict pathogenicity of tandem repeat loci

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Tandem repeat (TR) expansions are responsible for approximately 60 monogenic diseases.
• Identifying novel pathogenic TR expansions is key to reducing the diagnostic gap in genetic disorders.
• Interpreting the pathogenicity of newly discovered TR expansions presents a significant challenge.

Purpose of the Study:

• To develop and validate a machine learning tool for predicting the pathogenicity of tandem repeat expansions.
• To distinguish between pathogenic and benign TR expansions accurately.
• To aid in the discovery and interpretation of TR expansions in genetic disease research.

Main Methods:

• Development of RExPRT (Repeat EXpansion Pathogenicity pRediction Tool), a machine learning-based approach.
• Utilizing an ensemble method for classification of TR expansions.
• Evaluating the tool's performance using precision and recall metrics.

Main Results:

• The RExPRT tool, employing an ensemble approach, achieved an average precision of 93% and recall of 83% in classifying TR expansions.
• Demonstrated high accuracy in distinguishing pathogenic TR expansions from benign ones.
• The tool's performance indicates its utility in genetic research.

Conclusions:

• RExPRT is an effective machine learning tool for predicting the pathogenicity of tandem repeat expansions.
• The high precision of RExPRT is valuable for prioritizing candidate loci in large-scale discovery studies.
• This tool can accelerate the identification and diagnosis of genetic diseases caused by TR expansions.