• Home
  • Research Platforms

Research Platforms: The Susztak Lab Discovery Ecosystem

The Susztak Lab has engineered a synergistic set of research platforms that enable large-scale, hypothesis-driven discovery directly in human kidney disease. These platforms integrate deep molecular resolution with longitudinal clinical outcomes and AI-enabled synthesis, providing a foundation for all scientific programs in the laboratory.

Human Kidney Biobank: The Molecular Foundation

The Power: This platform overcomes the "scarcity bottleneck" in nephrology by providing one of the world's most deeply phenotyped human kidney repositories  (n>4,600).

  • Key Innovation: We integrate "Honest Broker" curated clinical data with multi-modal   preservation—including Fresh Frozen, FFPE, RNALater, and Primary Cell cultures. This enables the simultaneous study of kidney structure, cellular function, and underlying genetics from the same patient.
  • Scientific Reach: The biobank supports unbiased discovery across the entire molecular and structural spectrum: genomics, epigenomics, proteomics, metabolomics, and pathomics.
Human kidney biobank illustration highlighting more than 4,600 samples, integrated molecular data, and AI-powered discovery
Human Kidney Biobank: The Molecular Foundation. The biobank contains more than 4,600 samples and integrates honest broker–curated clinical data with fresh-frozen, FFPE, and RNAlater-preserved tissue, as well as primary cell cultures. It supports genomics, epigenomics, proteomics, metabolomics, and pathomics, enabling unbiased, AI-powered discovery.

 

TRIDENT: Longitudinal Precision for Diabetes

  • The Power: Transformative Research in Diabetic Nephropathy (TRIDENT) moves beyond cross-sectional "snapshots" to capture the dynamic molecular trajectory of Diabetic Kidney Disease (DKD).
  • Deep Phenotyping: This multi-center, prospective cohort follows individuals with diabetes, pairing longitudinal tissue collection with non-invasive biofluids (urine and plasma).
  • Translational Impact: TRIDENT enables high-resolution patient stratification and the discovery of biomarkers that predict therapeutic response and disease progression long before clinical decline occurs.
TRIDENT: Longitudinal Precision for Diabetes
TRIDENT: Longitudinal Precision for Diabetes. This multi-center prospective cohort pairs longitudinal kidney tissue collection with non-invasive biofluids, including urine and plasma, to map the molecular trajectory of diabetic kidney disease over time. These integrated data support high-resolution patient stratification, biomarker discovery, and prediction of disease progression and therapeutic response.

 

Penn Medicine Kidney Biobank: Health-System Scale Validation

  • The Power: This platform extends discovery beyond individual cohorts by providing health-system–scale breadth and population diversity.
  • EHR Integration: By linking kidney tissue and biofluids to Electronic Health Record (EHR) data across the entire Penn Medicine system, we enable large-scale outcome analyses and real-world clinical validation.
  • Generalizability: This scale ensures that molecular discoveries are robust and applicable across a broad, diverse patient population and various stages of kidney disease.
Penn Medicine Kidney Biobank: Health-System Scale Validation
Penn Medicine Kidney Biobank: Health-System Scale Validation. This platform links kidney tissue and biofluids with electronic health record data across the Penn Medicine health system, integrating diverse patient populations through a unified data platform. Its health-system scale enables large-scale outcome analyses, real-world validation of molecular discoveries, and identification of generalizable therapeutic targets.

 

NephroBase: The AI Analytical Engine

  • The Power: NephroBase is our kidney-specific computational and AI platform that serves as the analytical "brain," connecting clinical, imaging, and molecular data across all other platforms.
  • Next-Gen Capabilities:
    • Unified Latent Representations: Standardizing cell states across 70+ kidney cell types to enable seamless cross-platform integration.
    • Zero-Shot & Transfer Learning: Utilizing AI to apply knowledge from large-scale reference atlases to specific clinical contexts.
    • Predictive Modeling: Moving from description to automated prediction of disease progression, pathology quantification, and therapeutic target prioritization.
NephroBase: The AI Analytical Engine
NephroBase: The AI Analytical Engine. NephroBase integrates clinical, imaging, and molecular data through a kidney-specific computational and AI core. Unified representations across more than 70 kidney cell types, together with zero-shot and transfer learning, support prediction of disease progression, automated pathology quantification, and therapeutic target prioritization.