Improving newborn screening accuracy through genome sequencing, targeted metabolomics, and machine learning

Area of Science:

• Genomics
• Metabolomics
• Bioinformatics

Background:

• Newborn screening (NBS) using tandem mass spectrometry (MS/MS) faces challenges with false positives and diagnostic delays.
• Current methods necessitate confirmatory testing, impacting timely intervention for metabolic disorders.

Purpose of the Study:

• To evaluate the integration of genome sequencing, expanded metabolite profiling, and artificial intelligence/machine learning (AI/ML) for improved NBS accuracy.
• To enhance the precision of identifying true positive cases and reducing false positives in NBS.

Main Methods:

• Dried blood spots (DBS) from 119 screen-positive cases were analyzed.
• Genome sequencing identified variants in condition-related genes.
• An AI/ML classifier analyzed metabolomic data to differentiate true and false positives.

Main Results:

• Metabolomics with AI/ML achieved 100% sensitivity in detecting true positives.
• Genome sequencing reduced false positives by 98.8% but had lower sensitivity as a standalone test.
• Pathogenic variant carriers contributed to elevated false-positive rates, particularly in VLCADD cases.

Conclusions:

• Targeted metabolomics combined with AI/ML demonstrates high sensitivity for true positive identification in NBS.
• Genome sequencing effectively reduces false positives, but integration with metabolomic data is key for comprehensive NBS accuracy.
• Parental or prenatal carrier screening may further improve NBS accuracy by identifying at-risk individuals.