As advances in genome sequencing continue to expand the reach of prenatal genetic testing, more families are gaining insights into their unborn baby’s health. At the same time, these tests often reveal genetic variants that clinicians cannot yet confidently classify as harmful or harmless, leaving families without clear answers.

A new SickKids innovation aims to address this challenge by helping clinicians interpret uncertain genetic findings with greater confidence, potentially enabling earlier diagnoses and more informed decision-making for patients, families, and care teams.

Co-led by Rosanna Weksberg, Clinical Geneticist and Senior Associate Scientist, and Sanaa Choufani, Senior Research Associate in Genetics & Genome Biology, the team has developed a machine learning method that transforms disease-specific DNA methylation signatures, known as episignatures, into versatile diagnostic tools capable of working across multiple sample types.

Published in The American Journal of Human Genetics, the study demonstrated that blood-derived episignatures can be successfully adapted to identify patterns associated with Down syndrome across a range of prenatal and postnatal tissues. This work builds on the team’s development of EpigenCentral, a platform that uses DNA methylation signatures to help classify variants of uncertain significance as disease-causing or benign and represents an important step toward extending the clinical utility of episignatures. By overcoming a key limitation in prenatal testing, the approach opens new possibilities for applying episignatures across a broader range of sample types to improve accessibility and support more accurate diagnoses of rare neurodevelopmental disorders.