Sai Zhang
Yale University, CT·Genetics
Common diseases are shaped by many genetic variants. Each variant may act only in certain types of cells, which makes it hard to tell which variants truly cause disease and how.
- Fields
Genetics
Computer Science
Mathematics & Statistics
- Topics
single-cell genetics
single-cell genomics
deep learning
- Grants
- 1 known grant, through Feb 2030
What they work on
- Develop a sequence-based deep learning model, trained on single-cell and individual-level data, that predicts gene expression from DNA sequence and estimates how genetic variants alter gene activity in each cell type.
- Design a unified deep learning framework that combines single-cell reference data with genetic association summary statistics to rank disease-critical cell types and find cell-specific risk variants and genes.
- Build nonlinear, deep learning-based genetic risk scores that integrate single-cell annotations with common or rare variants, improving on linear polygenic risk scores for individual disease prediction.
More about the research
Huge datasets now exist, such as the UK Biobank, which links genetics to health records, and the Human Cell Atlas, which profiles individual cells across human tissues. However, the computational tools for genetic data and for single-cell data have developed separately. As a result, researchers struggle to connect a DNA variant to the specific cells and processes it affects.
Why it matters
Better tools to link genetic variants to the cells they affect could clarify what drives complex human diseases and point to new therapeutic targets. Single-cell-informed genetic risk scores could also improve predictions of individual disease risk. Openly shared software and data would let other researchers apply these methods to their own diseases of interest.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Vikas Pejaver
Icahn School of Medicine at Mount Sinai, NY·Genetics
Apply the team's previously developed variant impact predictors to large NIH Common Fund genomic data sets, such as Gabriella Miller Kids First and the Undiagnosed Disease Network, to identify known and predicted disease-associated variants.
Ariadne Letra
University of Pittsburgh, PA·Genetics
Determine the genetic architecture of familial cleft lip and palate (NSCLP) by combining linkage, case-control association and polygenic risk score analyses of whole genome sequences from 172 multigenerational families and additional datasets.
Martin Zhang
Carnegie Mellon University, PA·Genetics
Use multimodal single-cell and spatial genomics data to pinpoint, with high precision, the cell types and cellular contexts most important for diseases identified through genome-wide association studies (GWAS).
The school
Yale University: 73 labs on PI Pages, $1.5B a year on research, 438 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.