sPARK.

Screening · Pathology AI · Antigen · Repair · KRAS

Translational precision oncology for pancreatic cancer — connecting tumor genomics, blood-based biomarkers, immune profiling, and early-phase trials to move the right therapy to the right molecular subtype.

Explore Our Research Latest Publications

1,202Member Publications

NEJM+ Nature Medicine

5Research Pillars

2026 · Recent publications

Setidegrasib in Advanced Non–Small-Cell Lung Cancer and Pancreatic CancerPark et al. 2026NEJM

What We Do

Research Programs

Five interconnected sPARK pillars driving translational breakthroughs in pancreatic ductal adenocarcinoma (PDAC) — now expanding across gastrointestinal cancers through the GRIT initiative.

High-throughput Screening

Genetic and small-molecule screens — including CRISPR-based functional genomics — to find combination strategies, resistance mechanisms, and novel targets in PDAC.

Pathology AI

AI-enabled pathology and multimodal data integration to connect tissue phenotypes with therapeutic vulnerabilities in PDAC.

Neoantigens & Biologics

Pipelines for neoantigen discovery and biologic modalities targeting tumor-specific epitopes.

DNA Damage Repair & PARP-Immunotherapy

Combining DNA-damage-repair targeting with immunotherapy in BRCA/PALB2-mutated and HRD pancreatic cancer.

Nature Medicine 2026

KRAS-driven Vulnerabilities

Targeting KRAS G12D — the most common variant in PDAC — with first-in-class degraders and RAS(ON) inhibitors.

NEJM 2026

GRIT — GI RAS Innovative Therapeutics

A cross-tumor initiative extending RAS-directed precision therapy across gastrointestinal cancers — from pancreatic cancer to biliary tract, gastric/GEJ, appendiceal, and colorectal cancer.

Initiative

Biliary Tract Cancer

Extending our KRAS/RAS and precision-oncology work into biliary tract cancer — genomic determinants of chemotherapy benefit today, and RAS-directed trials next.

2026

All research programs →

A closer look / Spatial biology

Cellular context. Molecular detail.

Selected windows from real spatial transcriptomic analysis connect cellular neighborhoods, nuclear signal, and marker expression.

A cropped 800-micrometre spatial window of Segger cell boundaries, colored by provisional cell type, with a 100-micrometre scale bar. — 01
Cells in context
An internal spatial window, colored by exploratory cell-type annotation.

A cropped portion of an existing UMAP embedding colored by exploratory cell-type annotation; not a physical tissue map. — 02
Patterns in expression
A partial UMAP view, using the same cell-type colors as the spatial map.

A small DAPI nuclear microscopy window in lilac false color, with a 10-micrometre scale bar. — 03
Nuclear detail
Measured DAPI signal in a separate microscopic window, shown in false color.

Fibroblasts
Epithelial / ductal*
Macrophages
Endocrine-associated
B cells
Plasma cells
cDC
Mixed / ambiguous

*Epithelial / ductal malignancy is unresolved. Cell-type labels are exploratory. UMAP is an expression embedding; the spatial and DAPI panels show different locations and scales.

Marker gene dotplot for ten selected clusters, with a Ward dendrogram. Dot area represents RNA detection fraction; color represents relative expression within each gene. — Marker expression, connected.
Dot area shows the fraction of cells with detected RNA. Color shows gene-wise scaled log-transformed mean raw counts. The dendrogram groups selected clusters by marker expression, rather than depicting a cell lineage or evolutionary tree.
View full-size plot

Latest Work

Recent Publications

Clinical and translational work across KRAS-directed therapy, DNA-repair deficiency and immunotherapy, neoantigens, and biomarker-led treatment selection.