Jiyang Yu Lab

Signaling, epigenetic, and metabolic factors are crucial drivers of network rewiring, and likely therapeutic targets in cancer, but are difficult to identify as they are commonly altered by posttranslational modifications or other mechanisms, thus referred as “Hidden drivers”. We have developed the NetBID algorithm to infer hidden drivers by integrating transcriptomics with other omics data, which identified Hippo kinase MST1 as a hidden driver of CD8a⁺ dendritic cells for antitumor immunity. NetBID has been widely used in basic biology and oncology. For example, by using NetBID, we discovered that LCK is the hidden driver of unexpected dasatinib sensitivity in >40% of pediatric T-ALL and intra-tumoral heterogeneity of T-ALL is characterized by T-cell development stages conferring distinct therapeutic vulnerabilities to dasatinib and venetoclax driven by a dynamic LCK–BCL2 switch.

To decipher wiring and rewiring within and between individual cells and pinpoint underlying hidden drivers, we utilize single-cell and spatial omics technologies. We are developing a mutual information–based framework, scMINER, for clustering analysis, cell type–specific inference of intracellular networks, hidden drivers and network rewiring, and inference of intercellular communication networks. With spatial omics data, we are developing machine learning algorithms to refine cell-cell communication networks inferred from single-cell data and project hidden drivers into space. We have direct access to various single-cell and spatial omics technologies, including 10x Genomics (scRNA-seq, scATAC-seq, scMultiome, Visium), NanoString (GeoMx, CosMx), and CODEX.

Together with Drs. Hongbo Chi, Terrance Geiger, and Stephen Gottschalk, we are co-leading the St. Jude iTARGETS Blue-sky Initiative (Immuno-Oncology Target Identification via Systems Immunology). The goal is to identify and translate innovative targets in immuno-oncology, ultimately enabling and advancing curative therapies for pediatric cancers and other catastrophic diseases. In particular, we are reconstructing the tumor and tumor microenvironment interaction maps by integrating bulk, single-cell, and spatial omics technologies for novel immunotherapy target identification in pediatric cancers. We are also working to identify and target hidden drivers to overcome CAR-T resistance, improve CAR-T persistence, and develop potential combination strategies with immune checkpoint blockades in pediatric cancers, especially pediatric solid and brain tumors.