Many cancers are characterized by heterogeneity among individual tumors and between regions in a single tumor. The heterogeneity between regions, termed intra-tumor heterogeneity (ITH), has been shown to contribute to treatment failure and drug resistance through mechanisms, such as the expansion of pre-existing resistant subclones or immune suppression. By coupling single cell sequencing data (single cell RNA-seq, single cell DNA-seq, CyTOF and etc.) and in situ imaging data (H&E staining, CyCIF, and etc.), we are able to decode the cell populations in different regions of a single tumor, and identify new features that impact tumor growth in different therapeutic conditions. Such features may serve as new biomarkers for treatment response or new therapeutic targets in the drug development.

Mammalian genomes are organized into a hierarchy of local structures including topologically associating domains (TADs) and DNA loops. Disrupting the boundaries of such structures can lead to novel chromosomal interactions and ectopic long-range enhancer adoption, which interrupts key gene function. By using Hi-C, ChIA-PET and ChIP-seq data, we are able to precisely model the local structure of human genome and its impact in gene regulation. By using CRISPR screen with custom designed sgRNA library targeting the boundaries of such local structure, we could interrogate the mechanism of controlling gene expression through 3D genome organization.