Research Experience

• IPF niche analysis. Multi-omics analysis on large-scale datasets (>100M cells); applied deep learning to jointly model spatial coordinates and transcriptomics, detecting distinct tissue domains in VisiumHD and Xenium data.
• IPF polygenic risk scores & single-cell morphology. Bridging individual clinical background and genetics with spatial transcriptomics and cell morphology (H&E / IHC).

• Stem-cell-derived islets for Type 1 Diabetes. Collaboration with Vertex, Curio, and Harvard (Melton Lab). Ran differential expression, pathway enrichment, trajectory, and cell-cell communication analyses on single-cell data, integrating spatial data to assess distance effects on cell-type composition and interactions.
• CosMx kidney analyses. Collaboration with the University of Calgary (Chun Lab). Managed all computational analyses and built an accurate cell-type annotation pipeline for CosMx kidney tissue, generalizable to MERFISH and Xenium — yielding two papers (methodology, first author).
• CD4+ helper T-cell atlases. Developed an algorithm to predict CD4+ helper T-cell subtypes and built integrated atlases (Th2, Th17, Treg, Tfh, Th1) directly from raw matrices without Scanpy/Seurat. Th2 atlas accepted as a RECOMB 2024 poster; now in collaboration with Johnson&Johnson.
• Immune landscape in Classic Galactosemia. Spatial and single-cell analyses in collaboration with NIH scientists.
• Single-cell clonotypes (scTCR/BCR). Developed a clonotype-processing pipeline and published an open-source package for clonotype analysis and visualization.

• Back-end software engineer. BioTuring's product ecosystem — Talk2Data, BBrowserX, Lens
• Designed analysis pipelines. Spatial-omics technologies: Visium, GeoMx, Xenium, CosMx, MERFISH, CODEX, CyCIF, mIHC.
• Developed API packages. Bioflex, BioTuring SDK / BioTuringConnector.

• Curated single-cell datasets from public studies. Helped add ~1,000 studies in five months — then the world's largest curated single-cell database.