RefinedScience combines the deepest patient data in oncology, single-cell resolution, and real tissue models to give pharma partners the biological evidence they need—for target validation, patient selection, and clinical execution.
A joint venture between CU Anschutz and UCHealth.
In 2024, pharma spent $190 billion on R&D while Phase II success rates stood at 25%, with the composite probability of Phase I to approval at just 7%. The gap is largely biological: target validation that doesn't hold in patients, preclinical models that don't translate, and trial designs that enroll too broadly.
RefinedScience addresses those gaps directly. We build clinically-linked, single-cell datasets in specific oncology indications, validate biology in real patient tissue, and apply that depth to support target discovery, patient stratification, and clinical trial design.
We are a biology-first company providing what AI alone cannot: single-cell resolution of clinically annotated human tissue, with informatics and lab capabilities that improve the quality of biology at every stage of development.
The earlier we engage, the greater the compounding benefit across the full program arc. We are not a point solution for a single stage of development—we are a biology partner for the full journey.
Our value proposition isn't faster discovery—it's better-validated biology entering the clinic.
The RefinedScience platform combines unique institutional access to clinical data and patient tissue with AI analysis tools and biological validation.
Unified access to UCHealth longitudinal patient data, biobank samples, and EHR data all in one place—linked to biological samples across 7 oncology indications.
Proprietary software tools including scExploreR (single-cell data explorer), Marksman (target identification), Scout (AI co-scientist), and PULSE. 1,000+ AI-generated reports and 12 standardized protocols.
700+ assays developed. Functional drug testing on primary liquid tumors. Patient-derived organoids and spatial transcriptomics for solid tumors.
Biostatistics, data science, hematology, oncology, and precision medicine expertise—spanning bench to clinic. Joint venture of CU Anschutz and UCHealth with institutional access external vendors cannot replicate.
In October 2025, Verily announced a multi-year strategic collaboration with UCHealth, CU Anschutz, and RefinedScience to create AI-ready biomedical data pipelines and develop novel AI/ML models. An initial collaboration demonstrated 95%+ accuracy and 30× faster data extraction from AML patient records versus manual methods.
The RefinedScience Longitudinal AML dataset—covering 300+ venetoclax and azacitidine treated patients, with clinical and CITE-seq data—is now available on the Verily Exchange.
We work with both large pharma and emerging biopharma companies—at any stage of development, from target identification through clinical execution. Every engagement is shaped around the specific biology problem you need to solve.
The earlier we engage, the greater the compounding benefit across the full program arc. Biology validated at the discovery stage reduces cost and improves the design of every subsequent phase.
If you have a program in AML, MDS, PDAC, Ovarian, Esophageal, Glioblastoma, or DLBCL, we'd like to talk.
Partnerships@refinedscience.comEach engagement below reflects a specific problem we were asked to solve.
Full translational and clinical operations support for the Phase II study of cusatuzumab (anti-CD70) in combination with venetoclax and azacitidine in elderly patients with newly diagnosed AML ineligible for intensive chemotherapy.
Single-cell profiling of AML patients to identify molecular subgroups most likely to benefit from a partner compound. Enabled patient enrichment strategy for clinical development.
Preclinical evidence supporting a targeted AML therapy's efficacy and identification of predictive biomarkers. Single-cell resolution provided clarity on responding vs. non-responding populations.
RefinedScience supported ImCheck Therapeutics through real-world external dataset comparison and propensity-matched analysis. This work was included in ImCheck's FDA Breakthrough Therapy Designation application and contributed to the biological credibility that supported the company's acquisition by Ipsen.
General pharma R&D services, ADC discovery, or a discussion about a specific asset—we'll give you a direct answer on whether our data and capabilities are a fit.
Our platform generates commercial value through partnerships and through proprietary pipeline assets—starting with OncoVerity's Phase II program in AML.
We have built the AML datasets, primary tissue models, surface antigen mapping, and tumor-initiating cell models required for systematic ADC target discovery.
Multidisciplinary team bridging world-class academic research with commercial drug development expertise.
Our Values
RefinedScience supports pharma partners across three phases of drug development: Discovery, Translational Validation, and Clinical Development. Each engagement is defined by the specific biology problem you need to solve—not a fixed service catalog.
We are a biology-first company that brings proprietary data assets, analytical tools, and expert scientific staff to bear on your specific program. We engage where the biology problem is hardest—at the intersection of target validation, patient selection, and preclinical prediction.
Our institutional position—as a joint venture of CU Anschutz and UCHealth—provides unified access to longitudinal patient data, biobank samples, and EHR data all in one place, along with clinical infrastructure that cannot be replicated through commercial channels.
Systematic identification of surface antigens expressed on malignant cells and specifically on tumor-initiating cells—using CITE-seq surface protein mapping and functional validation.
Identify and validate molecular targets for small molecule intervention using pathway profiling, transcriptomic signatures, and resistance mechanism analysis.
Single-cell characterization of how tumors develop resistance to standard of care or investigational therapies. Design combination strategies to address surviving cell populations.
Confirm that your identified resistance mechanisms hold in real patient tissue across diverse patient populations. Move from hypothesis to validated biology before Phase I.
Identify synergistic combination partners based on single-cell pathway analysis and functional drug testing. Prioritize combinations with biological rationale, not empirical screening.
Functional drug testing on primary patient tissue—liquid tumors and patient-derived organoids with spatial transcriptomics for solid tumors. AI-informed lab-in-the-loop for efficacy prediction.
Single-cell profiling provided preclinical evidence supporting a targeted AML therapy's efficacy and identified predictive biomarkers for patient enrichment in clinical development.
View all partner work →Single-cell resolution to identify patient subpopulations most likely to respond. Molecular eligibility criteria development and enrichment models that reduce trial size while increasing probability of success.
Identify, develop, and validate biomarkers that differentiate responding from non-responding patients. Correlative science using multimodal data to provide biological context for clinical readouts.
Propensity-matched analysis using our real-world patient datasets to provide external comparators for single-arm studies. Used in the ImCheck FDA Breakthrough Therapy Designation application.
Simulate trial outcomes before enrollment. Test eligibility criteria, power assumptions, and stratification strategies using real patient populations.
Full clinical operations infrastructure for correlative science work. Delivered 500+ CSA-supporting assays for OncoVerity in under 5 months.
Companion diagnostic development using omics data to identify molecular signatures defining responding patient populations. Supports regulatory strategy and label differentiation.
Our indications are built disease by disease—each one combining clinical records, biobank tissue, and single-cell omics into a longitudinal, multi-modal resource.
Deepest indication. Longitudinal clinical data on 300+ patients with linked tissue. Venetoclax/azacitidine-treated patients with CITE-seq. Available on Verily Exchange.
Single-cell omics and clinical data in MDS—high unmet need with strong biological overlap with AML.
Single-cell profiling and spatial transcriptomics in one of oncology's hardest indications. Target discovery and combination identification focus.
Single-cell omics dataset including high-grade serous histology. Supports antibody-based target discovery and patient stratification.
Single-cell and clinical data in esophageal cancer. Supports target discovery and patient stratification research.
Multi-modal data in GBM. Single-cell resolution to characterize tumor microenvironment and identify subpopulation-specific targets.
Single-cell transcriptomics and clinical data in DLBCL for resistance profiling and patient stratification.
Describe your program and what you're trying to solve. We'll give you a direct answer on whether our data and capabilities are a fit.
Antibody-drug conjugate (ADC) target identification and validation for acute myeloid leukemia (AML)—unmatched clinical dataset depth, functional tumor-initiating cell (TIC) models, and single-cell multi-omics.
ADCs have demonstrated compelling clinical potential—but AML presents specific biological requirements. A target must be expressed on malignant cells and, critically, on tumor-initiating cells (TICs): the small, treatment-resistant subpopulation that drives relapse and death. It must be absent or low on healthy hematopoietic cells to avoid unacceptable toxicity.
Systematic ADC target evaluation in AML requires single-cell surface antigen mapping at the TIC level, functional primary tissue models with tissue validation, and longitudinal clinical data tracking how patients treated with related targets actually responded. RefinedScience has built that combination.
Our foundational ADC discovery platform is complete. We are seeking a limited number of partners to advance specific programs toward lead candidates.
Using CITE-seq protein measurement combined with single-cell transcriptomics, we map surface antigen expression across every major cell population in AML—at single-cell resolution. This produces a ranked list of candidate targets sorted by differential expression.
A target that is high on blasts but absent on TICs may not prevent relapse. Our platform maps target expression on tumor-initiating cells—the subpopulation most likely to drive disease recurrence—so candidates can be evaluated against the cells that matter most.
Every candidate target is evaluated against our longitudinal clinical dataset—how did patients whose tumors express this antigen respond to standard of care treatment? Clinical context eliminates biologically interesting but clinically irrelevant targets early.
Validated targets move to functional testing in primary patient tissue—not cell lines. We test ADC candidates for activity against malignant cells and TICs while evaluating their profile against normal hematopoietic populations.
We are seeking pharma partners who want to co-develop ADC programs in AML—bringing their ADC platform capabilities, antibody engineering, or clinical infrastructure to our target and disease biology.
Organizations with validated ADC technology seeking biologically differentiated targets in hematologic malignancies. We provide target, disease biology, and patient stratification.
Companies with existing AML programs looking to add a biologically validated ADC candidate. Our data can also support combination rationale.
Earlier-stage organizations building focused oncology ADC portfolios who need the disease biology and patient data foundation that takes years to build independently.
We are attending AACR and actively building our materials for BD conversations. If you have an ADC platform or are developing an AML program, reach out directly.