Current Research

Most diseases are complex, depending on multiple genetic, environmental and lifestyle factors. High throughput experiments facilitate data collection allowing some of that complexity to be captured. Different types of data provide varied, often complementary information about the same biological phenomenon. For example, genomic analyses of copy number alterations in cancer cells report on losses and amplifications of genomic regions, while genome wide RNA expression data provide evidence on the impact of genomic and environmental events on the internal wiring of the cell. In our lab we work on creating machine learning methods that make such data useful for clinical diagnosis.

Graphical models for variant aggregation

Cardiac Arrest Prediction

Similarity Network Fusion

Patient Heterogeneity

Drug Response Prediction

Dropout Feature Ranking of Deep Learning Models

Genetic Subphenotyping

Active miRNA Selection

More Publications