The next biotech platform may already exist in pieces.
The Venture Bench finds complementary technologies, inside one institution or across many, connects them around a shared clinical problem, and helps test whether they deliver more predictive value together than apart.
For universities, academic medical centers, research institutes, and biotechnology companies with technologies that may be stronger in combination.
Innovation is abundant. It is also fragmented.
Promising diagnostic, therapeutic, computational, and platform technologies are usually evaluated, protected, and commercialized one asset at a time. That pathway matters, but it can hide the value that sits between technologies.
Silos inside the same institution
Investigators in one university or medical center may work on the same disease from different points in the standard of care: imaging, pathology, genomics, proteomics, metabolomics, liquid biopsy, or computational prediction. Separated by department, funding, or data environment, the value of combining their work can go unrecognized.
Silos between institutions
Complementary technologies addressing the same disease, pathway, clinical decision, or patient population may never be evaluated together simply because they originate in different organizations: another university, a startup, a hospital dataset, or an engineering lab.
What other technology, inside this institution or somewhere else, could make this one substantially more useful? Traditional scouting asks whether a technology is good. The Venture Bench also asks what it could become in combination.
The Venture Bench connected innovation framework.
A structured approach for identifying complementary technologies across disciplines and organizations, connecting them around a shared problem, and testing whether integration creates measurable value. We refer to this working framework as Combineering.
Discovery
Look across investigators, departments, portfolios, institutions, and disciplines for technologies aimed at the same indication.
Complementarity
Confirm that each technology measures a meaningfully different dimension of the same biological or clinical problem.
Integration
Determine whether the technologies or their data can be combined in a scientifically and operationally credible way.
Validation
Test whether the integrated approach provides incremental value over each individual component.
The output is a map of technology relationships, not simply a collection of assets.
Different modalities, different signals.
The strongest opportunities pair technologies that capture complementary or orthogonal information. If two technologies measure essentially the same thing, combining them adds little.
We describe the relationship between modalities as potentially synergistic. What must be demonstrated is incremental value: whether the combined model measurably outperforms each component on its own.
| Technology modality | Illustrative information provided |
|---|---|
| Medical imaging | Structure, phenotype, spatial characteristics |
| Liquid biopsy | Molecular or circulating biological signals |
| Genomics | Genetic architecture |
| Proteomics | Protein-level biology |
| Metabolomics / metabolic profiling | Metabolite-level physiology and biochemical state |
| Digital biomarkers | Continuous behavioral or physiological information |
| Clinical data | Patient risk and disease context |
| Artificial intelligence | Relationships across multiple data types |
Combining is a hypothesis. Improvement must be proven.
The framework never assumes two technologies perform better together. Each engagement compares the integrated model with every individual component and asks whether any statistical gain changes a real clinical or operational decision.
Measures we consider beyond AUC
- Sensitivity and specificity
- Positive and negative predictive value
- Calibration
- Patient stratification
- Clinical decision curves
- Avoided procedures or interventions
- Treatment selection
- Health-economic outcomes
What the principle looks like in published research
Predicting malignant pathology in pancreatic cysts (IPMNs). Area under the ROC curve by model.
For diagnostics, validation can be staged through pre-agreed performance gates, from single-site validation to multi-site replication. See how that works in the diagnostic de-risking pathway.
Where connected innovation applies.
These examples are intentionally generic. Several describe active areas of scientific investigation rather than settled mechanisms.
Phenotype and molecular biology, read together
Independently developed technologies can provide different information along the same diagnosis and treatment continuum.
Imaging / radiomics
- Lesion morphology
- Spatial characteristics
- Tissue phenotype and heterogeneity
- Quantitative radiographic features
Liquid biopsy
- Molecular signals of disease presence
- Biological aggressiveness
- Progression
- Treatment response or resistance
Integrated predictive algorithm
Probability of clinically significant disease
Validation question: does the combined model add predictive value beyond imaging alone and liquid biopsy alone?
Better-informed surveillance and intervention decisions
Structural imaging and molecular information may be complementary at the risk-assessment step of pancreatic cyst management.
Imaging / radiomics
- Cyst morphology
- Anatomical characteristics
- Structural heterogeneity
- Imaging features associated with risk
Liquid biopsy
- Molecular biomarkers of disease
- Progression
- Biological aggressiveness
- Malignant transformation
Multimodal algorithm
Risk stratification: surveillance or surgical consideration
Validation question: does the integrated system add meaningful predictive and clinical value beyond its individual components?
Where the brain is changing, and why
Neuroimaging may characterize where and how the brain is changing, while a blood-based assay may reflect the underlying molecular pathology.
Neuroimaging
- Regional brain volume
- Structural change
- Functional connectivity
- Metabolic patterns
Blood-based biomarker
- Molecular signals associated with neurodegeneration
Integrated predictive algorithm
Probability of disease or progression
Validation question: how does the integrated model compare with imaging alone and the blood biomarker alone?
Connecting gastroenterology and neurology
Constipation can precede the motor symptoms of Parkinson disease, and research is examining links among gut dysbiosis, intestinal permeability, inflammation, and misfolded α-synuclein. A proposed gut-first model remains under active investigation rather than a settled causal sequence.
Gastrointestinal modalities
- GI motility and constipation phenotyping
- Microbiome and metabolomic profiling
- Intestinal-barrier biomarkers
- Enteric α-synuclein assays
Neurological modalities
- Autonomic measures
- Neuroimaging
- Blood or CSF biomarkers
- Clinical progression data
Integrated model
Gut-dominant phenotype and earlier risk stratification
Validation question: can an integrated model predict neurodegenerative progression better than any single modality?
Cardiovascular, neurovascular, and Alzheimer’s biology
Alzheimer’s disease and vascular health intersect through shared risk factors and the neurovascular system. Cerebral blood flow, blood-brain barrier integrity, and vascular health may interact with amyloid and tau biology.
Cardiovascular phenotyping
- Blood pressure and arterial stiffness
- Cardiac function
- Vascular risk burden
- Perfusion
Neuro biomarkers
- Neurovascular imaging and MRI
- Amyloid or tau imaging
- Fluid biomarkers
- Computational prediction
Integrated model
Prediction of cognitive decline and mixed vascular-neurodegenerative risk
Validation question: does combining these signals improve prediction of cognitive decline?
Why people respond differently to the same food
One technology describes what happened after a person ate. The other may explain why that response differs from someone else’s.
Physiological monitoring
- Post-meal glucose response
- Metabolic patterns
- Activity and sleep
- Longitudinal changes after food intake
Microbiome or molecular profiling
- Gut microbial composition
- Microbial functional signatures
- Metabolites
- Molecular biomarkers
Personalized prediction algorithm
Predicted response to a nutritional intervention
Validation question: does the integrated model, with dietary information, predict individual response better than either source alone?
From a defined problem to a validated, scalable solution.
Eleven steps take a single technology through discovery, integration, validation, and translation.
Define
Identify the specific clinical, biological, or commercial problem.
Deconstruct
Determine what the technology measures, what it predicts, and where its limits are.
Map the standard of care
Locate the technology in the clinical workflow, from screening to follow-up.
Map complementary modalities
Identify dimensions that could add information at the same or adjacent decision points.
Scout
Search within the institution and across external institutions, companies, and portfolios.
Match
Evaluate scientific compatibility between candidate technologies.
Connect
Bring together investigators, departments, technology owners, and development partners.
Integrate
Develop a strategy for combining the technologies or their datasets.
Validate
Compare the integrated system with each individual component.
Translate
Evaluate clinical, regulatory, intellectual property, commercial, and implementation pathways.
Scale
Build the partnerships and strategy to advance the integrated solution.
From asset evaluation to relationship evaluation.
Our discovery matrix adds a second question to every conventional one.
| Conventional technology evaluation | The Venture Bench evaluation |
|---|---|
| What is the current standard of care? | Where does this technology fit, and which adjacent modalities or decision points could create a stronger integrated solution? |
| What does the technology do? | What does the technology not yet tell us? |
| Is the technology novel? | What complementary technology could increase its value? |
| What is its standalone performance? | What independent signal could improve its prediction? |
| Who owns the technology? | Who, within or outside the institution, has a complementary technology? |
| What is its market? | Could integration create a new market opportunity? |
| Can it be licensed? | Can it become part of a larger solution? |
| What data does it generate? | What other dataset could make that information more useful? |
| Who might commercialize it? | Which partners are needed to create the integrated platform? |
The value of a technology may lie not only in what it does alone, but in its potential combinations.
A new role for technology portfolios
Traditional portfolio management asks which technologies should be advanced. The Venture Bench adds a second question: which technologies should be connected? A portfolio of hundreds of early-stage technologies also holds a far larger number of possible relationships, across investigators, institutions, and startup platforms.
Intellectual property and commercial opportunity
Integration may create new innovation in multimodal algorithms, integrated diagnostic methods, clinical workflows, and data-fusion methods. Patentability, ownership, inventorship, freedom to operate, licensing, data rights, and regulatory questions require case-specific analysis, so we raise them early and build them into the plan.
Connected innovation services.
We operate between discovery and translation, and between individual technologies and integrated solutions.
Technology intelligence
Understanding what a technology does, the problem it addresses, where it sits in the standard of care, and what could strengthen it.
Modality relationship discovery
Identifying high-value relationships across imaging, pathology, omics, liquid biopsy, digital measures, computation, and clinical data.
Intra-institutional mapping
Finding investigators, departments, and technologies within one organization that address a shared indication through complementary modalities.
Inter-institutional scouting
Searching other universities, medical centers, startups, companies, and research institutes for complementary technologies.
Technology matching
Identifying scientifically plausible relationships between otherwise independent assets.
Collaboration architecture
Connecting the investigators, institutions, companies, clinicians, and stakeholders needed to evaluate a combination.
Validation strategy
Defining how the combined system should be tested against its components to demonstrate meaningful incremental value.
Many organizations ask what has been invented. We ask what these inventions could become together.
A team that looks between the assets.
The Venture Bench team members, informally called Combineers, search for meaningful connections between technologies that were developed independently.
The work requires curiosity across disciplines, comfort with technologies outside one’s own field, a working knowledge of translational pathways and the standard of care, and a willingness to look across laboratory, departmental, and institutional boundaries.
Of every technology we evaluate.
- What problem is this technology trying to solve?
- Where does it fit in the current standard of care?
- What information does it provide, and what is missing?
- What limits its current predictive or clinical performance?
- What other modality could provide an independent signal?
- Does that technology exist elsewhere in the institution, or at another one?
- Could the two datasets be integrated, and would the combination improve prediction?
- Would that improvement change a clinical or operational decision?
The pieces may be closer than you think.
One may sit in an imaging lab and another in a pathology group down the hall. Or the missing component may be at another university, a startup, or a hospital dataset. We help you find it and test whether it belongs.
- Siloed to connected technologies
- Connected technologies to integrated data
- Integrated data to validated prediction
- Validated prediction to translational value