AI expands the design space
Scientific models can assist with protein sequence, structure, function, experimental design and interpretation.
Long-term product vision · August 2026
From a digital hypothesis to verified biological reality
A closed biotechnology Design–Build–Test–Learn cycle—with method validation and evidence governance—without requiring each user to own or operate a laboratory.
Due Science’s long-term goal is to become the evidence-execution layer between AI-assisted biotechnology design and physical biology: an independent orchestration layer that turns human- or AI-generated hypotheses into traceable experimental evidence and returns structured results to the next design iteration.
Long-term category · Biotech Verification-as-a-Service
Founding pilotWritten scoping intake open
Digital hypothesisExperimental evidence
01DESIGN✓LOCK02EXECUTE
Evidence DocketPROTOCOL · DATA · REVIEW
STRUCTURED RESULT → NEXT DESIGN ITERATION
The infrastructure opportunity
Toward creating and testing biotechnology without owning laboratory infrastructure.
Like cloud infrastructure makes it possible to run software without owning servers, Due Science aims to make it possible to create and test biotechnology without owning a laboratory or directly operating its equipment.
Artificial intelligence (AI), including specialised scientific models and large language models (LLMs), is expanding the range of molecules, proteins, protocols and biological hypotheses that people can design in a digital environment.
AI-assisted software development becomes productive when generated code can be executed, tested and inspected. Independent testing agents and an orchestration harness turn each result into the next development cycle. Actual execution remains the common reference point.
Biotechnology is moving toward a similar operating model. Digital design can propose a candidate. Physical biology determines how that candidate behaves. The next infrastructure opportunity is a traceable connection between those two environments.
Verificationasks whether an experiment supports a specific claim.
Validationestablishes whether the method can consistently measure the intended property.
Why now
Generation is scaling. Governed experimental learning can scale with it.
Robotic laboratories structure execution
Automated workflows can produce instrument files, images, plate-level data, timestamps and run metadata in consistent digital formats.
Biotech decisions need bounded evidence
An investment, licence, option, financing tranche or portfolio milestone may depend on one carefully scoped experiment.
One controlled learning loop
Design–Build–Test–Learn, with validation and evidence governance.
Design–Build–Test–Learn (DBTL) is the recurring cycle used in engineering biology. Due Science is designed to connect each stage through one traceable evidence record.
- D
Design
A person or specialised model formulates a hypothesis and a digital candidate.
- B
Build
An external laboratory converts the design into physical material or a cellular test article.
- T
Test
A prespecified robotic workflow executes the experiment with defined controls and repeats.
- V
Validate
The method, raw instrument data, execution record, deviations and analysis are reviewed against agreed criteria.
- L
Learn
The structured result returns to the digital environment and informs the next candidate or experiment.
The objective is a shorter, more accessible and more reproducible path from hypothesis to physical learning.
The product path
One verification. A reusable evidence lifecycle. A platform category.
Evidence Docket
The versioned record of one verification: claim, protocol, criteria, sample and data provenance, execution metadata, deviations, raw instrument data, analysis and conclusion boundaries.
Evidence Passport
The planned lifecycle extension will connect successive Evidence Dockets, authorised access, disclosures, reliance rights, milestones and repeat tests.
Biotech Verification-as-a-Service
Programmatic access to distributed robotic workflows for people, biotechnology teams and scientific agent systems—artificial-intelligence systems that can use tools and carry out research tasks.
BrandDue Science
Buyer categoryIndependent Experimental Diligence for Biotech Transactions
Investor categoryAI-Native Evidence Infrastructure for Life-Science Transactions — infrastructure designed from the outset around artificial intelligence (AI)
Long-term categoryBiotech Verification-as-a-Service
ProductEvidence Docket (versioned evidence record) · planned lifecycle layer: Evidence Passport
TechnologyArtificial-intelligence-assisted orchestration · expert governance · independent laboratory execution
Founding Verification Pilot · Written scoping intake open
Independent Experimental Diligence for Biotech Transactions
Due Science provides a live planning-range calculator and written request route for buyers considering one public, decision-critical scientific claim before an investment, licensing or portfolio decision. An accepted project would convert that question into a locked experiment, identify and qualify a compatible independent robotic laboratory, coordinate contracted execution and assemble the result into a versioned Evidence Docket.
The narrow founding-pilot scope covers simple proteins, cytokines—proteins that signal between cells—enzymes and peptides, with one public reference, one primary quantitative endpoint and an investment, licensing or milestone decision context. Regenerative medicine and longevity are preferred initial application areas.
- 01Decision Question
Define the claim, decision owner, deadline and available material.
- 02Verification Design Lock
Fix the primary measurement, controls, repeats, result threshold and permitted interpretation before execution.
- 03Laboratory Routing
Compare compatible sites by technical fit, total cost, timing, data rights and contract terms.
- 04Physical Execution and Review
Collect raw instrument data and execution records, followed by accountable scientific and statistical review.
- 05Evidence Docket
Deliver the protocol, documented data origin and handling history, deviations, data, analysis and a conclusion limited to the prespecified claim as one versioned record.
Physical execution begins after documented qualification of the specific laboratory site and workflow version, agreement of the scope and price, and execution of the applicable contracts. Due Science then uses the internal Qualified Robotic Laboratory (QRL) designation for that site-and-workflow combination; the designation is distinct from regulatory accreditation.
Open the founding pilotAvailable now
A live written-scoping route for a narrow founding pilot.
The public pilot provides category-level testability scoping, an indicative cost-and-timeline range and a controlled written request. No scientific files or claim URL are collected on the website. A compatible independent laboratory is identified, conflict-screened, qualified, quoted and contracted separately for each accepted project.
The calculator and written request route are live. They do not place an order or reserve laboratory capacity. Physical execution becomes available only after project-specific feasibility review, qualification, quotation and contracting.- 01Category-level scope and planning range
- 02Written preliminary testability response
- 03Controlled request for one public reference
- 04Separate scope, quotation and project contract
- 05Project-specific laboratory qualification and potential execution
The economics of iteration
Make each well-documented learning cycle easier to repeat.
The central economic unit is the cost and time of one well-documented learning cycle. Reusable experimental templates, controls, contracts, data formats and laboratory routing can make repeated testing progressively easier.
Both supportive and non-supportive results become structured knowledge. Teams can explore more candidates, confirm the strongest findings and direct resources using observed biological evidence.
A clear interface can also broaden participation. Clinicians, chemists, engineers, data scientists, computational biologists, disease foundations and entrepreneurs can formulate a question in the language of their field while the platform translates it into a governed experimental workflow.
A future regulated pathway
Evidence infrastructure for individualized medicine.
Precision medicine selects treatment using patient characteristics. An individualized, or N-of-1, therapy is developed for one person or a very small group.
Over time, the same architecture may support a Patient-Specific Evidence Docket covering identity, purity, manufacturing quality, functional activity, relevant risk indicators and provenance for a particular therapeutic version.
This is a future regulated pathway, distinct from the current transaction-diligence service. It would operate through licensed clinical laboratories and healthcare partners, with jurisdiction-specific privacy, consent, quality, regulatory and medical governance. A Patient-Specific Evidence Docket would contribute structured evidence within that wider system; clinical safety and effectiveness would remain determined through regulated development, clinical evidence, regulatory review and licensed-clinician judgment.
Trust is part of the product
The operating standard is designed to govern evidence before, during and after execution.
- 01
Hypotheses, controls and acceptance criteria are fixed before execution.
- 02
Funding and potential conflicts are disclosed; fees are independent of experimental outcome.
- 03
Samples, native data, instrument records, manual actions, deviations and analytical versions remain traceable.
- 04
Scientific, statistical, legal, medical and biosafety expertise enters at defined governance points.
- 05
Laboratory qualification is specific to the site, method and workflow version.
Share your perspective
Help shape the evidence layer for AI-native biotechnology.
We welcome concise written feedback from life-science funds, venture studios, diligence teams, biotechnology and pharmaceutical licensing professionals, robotic laboratories and scientific experts.
- Which buyer has the clearest need?
- Which bounded experiment could change a decision?
- What would make its Evidence Docket credible and reusable?