About

Anna Goldenberg develops machine learning methods that combine diverse sets of biological and phenotypic measurements to refine the understanding of complex human diseases, identify the best treatments and individual patient outcomes, and guide decisions to improve the quality of life for patients.

Her data integration method – Similarity Network Fusion (SNF) – was the first to integrate patient data (omics, imaging, etc.) using patient networks. It improved survival outcome predictions in five different cancers.

More than two-thirds of mental health issues have their onset during childhood or adolescence. Identifying children who are at risk for mental illness later in life, and predicting the type of illness, is not easy. There are no blood or genetic ‘tests’; instead, psychiatrists and other health professionals use an individual’s personal and family clinical history to note persistent symptoms and make a diagnosis. Goldenberg believes that by modelling and combining trajectories of social-emotional and neurocognitive measures over time, and using genome-wide genetic and DNA methylation data, it is possible to identify clinically relevant subtypes of childhood and adolescent development and also the corresponding biomarkers that predict traits underlying psychiatric disorders.

The easy-to-use computational tools developed to ascertain longitudinal subtypes and their associated biomarkers will be made available to the larger research community. The findings of Goldenberg and her team will implicate both short- and long-term predictors and biological contributors of psychopathology, helping to unravel the complex relationship between genomics and social-emotional development. Hopefully, these findings will also generate candidate biological targets for novel treatments.