NCI-supported researchers leveraged machine learning (ML) models to understand the specific mutations that initiate Clonal Hematopoiesis (CH)—a condition where genetically mutated hematopoietic stem cells multiply faster than usual, potentially leading to blood cancers and other health issues. By following their approach, you can use similar ML techniques and software to conduct your own analyses.

To replicate or extend their findings, you’ll need a large set of high-quality blood somatic mutations to train these ML models. Researchers achieved this by using samples from over 33,000 patients with cancer from three cancer genomics cohorts. Building on methods previously applied to identify cancer driver mutations, they trained their models on features that distinguish the blood somatic mutations observed in CH genes across donors from those expected to arise under neutral mutagenesis. With these tools and data sets, you can gain deeper insights into CH and its implications.

As researcher, Dr. Abel Gonzalez-Perez shared, “These models outperform expert-curated rules based on prior knowledge of the function of these genes. This suggests machine learning provides a more systematic and unbiased method for identifying CH driver mutations.”

He added, “We propose that these models support the accurate identification of CH across healthy individuals and imply the potential for early CH detection and intervention in seemingly healthy individuals.”