Machine Learning

Machine learning (ML), a branch of artificial intelligence, enables computers to learn from data and identify patterns that support predictions and decision-making. In genomics, ML models are increasingly being used to interpret complex biological data such as DNA, RNA, and protein sequences. These approaches can uncover hidden relationships between genetic variants and disease, predict which mutations may be pathogenic and reveal mechanisms that underscore neurological and rare disorders.

In neurological genomics, ML methods are already reshaping how scientists detect disease risk, identify biomarkers and classify subtypes of complex conditions. By combining biological knowledge with data-driven discovery, machine learning provides a powerful framework to move from large-scale sequencing data to meaningful clinical insights.

Why This Research Matters

At COIN, we apply machine learning to improve genetic diagnostics for patients with neurological and neuromuscular disorders, particularly those whose genetic causes remain unexplained after standard testing. Our team is developing ML-based models to more accurately detect exon-level copy number variants (CNVs) from whole-exome sequencing (WES) data, an area where traditional methods often fall short.

While short- and long-read whole-genome sequencing (WGS) have improved CNV detection globally, these approaches remain expensive and resource-intensive, limiting their accessibility in many African settings. WES, on the other hand, is more affordable but still challenged by technical constraints such as low precision, high false positive rates, and dependence on large, high-quality reference datasets. By applying ML to distinguish true CNVs from background noise, our research aims to overcome these barriers; enhancing the diagnostic yield of WES while making it a more powerful tool for rare disease genomics in Africa.

Developing cost-effective, high-accuracy CNV detection methods is particularly impactful in resource-limited contexts. It allows local research and clinical programs to perform advanced genetic analysis without relying on costly infrastructure, supporting earlier diagnoses and enabling discoveries that reflect Africa’s unique genomic diversity.

Looking Ahead

This work has the potential to transform rare disease diagnostics by increasing the accuracy and reach of exome-based testing. Improved detection of disease-causing variants can lead to earlier diagnoses, more informed genetic counselling and personalised treatment strategies. As these ML-driven tools evolve, they will help ensure that precision medicine advances are equitable and relevant, for Africa and the world.