Validate a novel target-disease hypothesis with a BVCT before committing to a Target Candidate Profile.
Configure the BVCT
This chapter covers Part B of the FIC Targets workflow: taking the regulated hypothesis you confirmed in the Generator and validating it with a BioinvestGPT Virtual Clinical Trial (BVCT). Part A — exploring the knowledge graph and annotating up/down-regulated targets — is covered separately; this chapter picks up the moment you land in the Validator.
What carried over: When you click "Confirm All Hypotheses" in the Generator, the platform packages your disease, regulated targets, and supporting literature and navigates you to the Validator. The disease and therapy fields are already filled in — you do not re-enter them here.
8
Configure the BVCT
The Validator opens on the RoutePlanner with your disease and therapy pre-filled from the Generator. Set the two routing controls that define what kind of trial you are simulating:
Control
Options
What it means
BVCT Type
ONGOING / PLANNING
ONGOING aligns the simulation to a trial already running; PLANNING designs a trial from scratch.
Purpose
VALIDATION / DISCOVERY
VALIDATION tests a specific hypothesis you already hold; DISCOVERY explores the response surface more broadly.
Optionally paste an NCT ID to align the BVCT with a real registered clinical trial — useful when you want the virtual trial to mirror a competitor's protocol or an ongoing study.
Tip: For a first-in-class target you are still derisking, the typical setting is BVCT Type = PLANNING and Purpose = VALIDATION — you are testing whether the hypothesis holds before you build the trial.
Before you click Continue
Confirm that the pre-filled disease and therapy match the hypothesis you authored in the Generator. If anything looks wrong, return to the Generator rather than editing here — the regulated targets and literature travel as a package, and re-entering only the surface fields can desynchronise them from the underlying hypothesis.
Patient stratification
9
Patient stratification
Click "Continue" to generate patient strata. BVCT synthesizes 40+ de novo virtual patient subgroups from disease biology, your target mechanism, and the trial design — no real patient data is required. Each subgroup combination receives its own predicted effect size from causal simulation.
Because the strata are derived from the hypothesis itself, they reflect the specific up/down-regulation you encoded in the Generator: the subgroups most sensitive to your target mechanism surface with the largest predicted effects.
Why de novo strata matter for a first-in-class target
A genuinely novel target has, by definition, no prior trial history to mine. De novo synthesis lets you stratify around a mechanism that has never been dosed in patients — the simulation reasons from biology rather than retrofitting historical cohorts. This is the core of derisking before you commit to a Target Candidate Profile.
Tip: Scan the strata for the subgroups with the strongest predicted effect size — these are your enrichment candidates. A target that only moves a narrow subgroup is still a valid hypothesis, but it points toward a biomarker-enriched trial rather than an all-comers design.
Eligibility criteria
10
Eligibility criteria
Standard ICH-compliant inclusion / exclusion criteria are auto-generated for the trial, including the special-population gates regulators expect to see:
Adequate organ function — hematologic, hepatic, and renal thresholds
Age thresholds — 18+ for adult indications, or 12+ where a pediatric indication applies
Pregnancy / lactation exclusion
How eligibility interacts with your strata
The eligibility criteria define the population the protocol can enroll; the strata from the previous step define how that population is predicted to respond. Read them together: if your strongest-responding subgroup would be excluded by a standard organ-function or age gate, that is a design signal to widen the criteria or to reposition the indication.
Note: Eligibility criteria are auto-generated to a conventional ICH baseline. They are a starting point for protocol design, not a regulatory submission — treat them as the draft a clinical operations team would refine.
Review protocol cards
11
Review protocol cards
Each patient stratum generates its own downloadable protocol card. Every card shows the four numbers that decide whether a stratum is worth running:
Field
What it tells you
Stratum name
The virtual subgroup this protocol targets
Predicted effect size
The Hazard Ratio the causal simulation predicts for this stratum
Required sample size
Patients needed to power the trial at that effect size
Trial duration
Estimated time to read out
Two actions sit on each card:
Download — export the protocol card as a PDF for circulation or filing.
Verify — mark the protocol as reviewed. Only verified protocols are eligible for transmission in the next step.
Tip: A strong predicted effect size paired with a small required sample size is the ideal combination — it means the hypothesis is both biologically plausible and cheap to test. Sort your attention by that pairing.
Reading effect size as a derisking signal
The predicted Hazard Ratio is the quantitative core of the derisking exercise. A favorable HR concentrated in a coherent, enrollable stratum is the evidence that turns a hypothesis into a Target Candidate Profile worth pursuing. Verify the cards that pass your bar; leave the rest unverified so they do not enter the transmitted package.
Approve and transmit
12
Approve and transmit
Once you have verified the protocols you want to advance, click "Approve Verified BVCT Protocols". The platform transmits the protocol package and confirms with "Transmission Finalized".
Only the protocols you marked Verify in the previous step are included — approval acts on the verified set, so unverified strata are left out of the transmitted package.
What "Transmission Finalized" means
Transmission packages your approved, verified protocols as the validated output of this BVCT. From here the hypothesis has moved from an unproven idea on the knowledge graph to a quantified, stratified, protocol-backed candidate — the basis for committing to a Target Candidate Profile.
Decision support, not a guarantee: A BVCT verdict is a model-based decision-support analysis, not investment advice and not a regulatory endorsement. Use it to prioritise and design — pair it with wet-lab and clinical validation before committing resources.
You can audit how BVCT predictions have held up against real outcomes on the prospective track record at data.bioinvestgpt.com.
Where to go next
To visualise the rationale behind a transmitted verdict back on the 3D knowledge graph, use the "Visualize Rationale" action in the BVCT result summary — it carries the trial drug, indication, and predicted effect size back into the Generator. To open the Validator directly for a future hypothesis, go to bvct.bioinvestgpt.com/#/validator.
BioinvestGPT BVCT Platform User Guide — Clinically Derisk FIC Targets (HYPOTHESIS). BVCT outputs are model-based decision-support analyses, not investment advice. Prospective track record: data.bioinvestgpt.com.
