Our group develops combined computational and experimental approaches to the discovery of novel biology and human therapeutics.    We create interpretable computational models that we train and validate with experimental evidence.   With collaborators, we apply these models to problems in experiment design, developmental biology, gene regulation, immunology, genomics, and human therapeutics.   We typically evaluate our models and discover new biology with multiplexed high-throughput experimental studies that produce data from populations of cells and single cells.

A continuing challenge is our incomplete knowledge of biological systems that leads to model uncertainty.   An active area of study in our group is the generation of appropriate uncertainty metrics for models, and how they can guide experiment design to improve model accuracy.   We use conventional large scale linear and non-linear models, Bayesian methods, and deep learning approaches in our computational thinking.

Our current specific biology focus areas include motor neuron development, single-cell perturbation studies, the regulation of chromatin accessibility, the regulatory genome, antibody design, and peptide presentation by major histocompatibility complex (MHC) proteins.