To capture the genome in 4D, we are improving on existing genomics techniques (Hi-C, RNA-seq), including in the realm of single cell analysis, and are developing new experimental and analytic techniques in imaging (e.g. oligo-DNA and RNA FISH). Coupling these techniques with mathematical analysis reveals natural genome dynamics and biological rhythms, including across the cell cycle and enables building data-guided algorithms that improve and generate new methods for cellular reprogramming. Our long-term goal is to understand the controllability of the 4D genome, and develop strategies for direct reprogramming of normal and abnormal cells into desired targets.

A mathematical theory of Learning guided by the Immune System. Our hope is that by gaining a better understanding of the immune system and the exact way that it identifies and counteracts antigens, we will be able to improve or possibly even create novel forms of machine learning. This will be done by formulating the mathematical framework of learning in the context of the immune system which will be generalizable to any other setting.

Direct Cellular Reprogramming Algorithm

We have developed an algorithm for the prediction of transcription factors necessary for direct cellular reprogramming - from any cell to any cell. This algorithm uses time series genomic data, both architecture and function, and control theory to determine the optimal inputs for controlling cell fate.

Simultaneous DNA and RNA oligo-FISH

DNA and RNA FISH imaging of CCNY during cell cycle progression.

Dynamic Genome Imaging

We have developed an objective algorithm for high-throughput, time-series FISH imaging, with canonical framework detection for gene positioning.

Single Cell Hi-C

Heat plots of single cell Hi-C maps from different cells (left 3) and a bulk Hi-C map (right). Green arrows in the Cell 1 and Bulk maps indicate chromosomal abnormalities seen in this cancer cell line.