# Galen pre-seed investor page Galen is the virtual-cell company. Galen builds causal virtual-cell models intended to predict how living cells respond to intervention, quantify confidence, and choose the next experiment most likely to teach the model. Disclosure: This packet contains the public investor-page content. Approved investors receive detailed diligence materials, protocol details, round materials, and private strategy through the investor data room. Source: https://preseed.usegalen.com Public Markdown: https://preseed.usegalen.com/llms.txt Public JSON: https://preseed.usegalen.com/investor-brief.json AI Diligence Prompt: https://preseed.usegalen.com/ai-diligence-prompt.txt Partner Memo: https://preseed.usegalen.com/partner-memo.md Proof Ledger: https://preseed.usegalen.com/proof-ledger.md Diligence Questions: https://preseed.usegalen.com/diligence-questions.md Intro Email: https://preseed.usegalen.com/intro-email.txt ## Forwardable Intro Galen is building causal virtual-cell models that predict how living cells respond to intervention, quantify confidence, and choose the next experiment most likely to teach the model. The pre-seed round funds the real human-cell proof loop: predictions committed before results, model-selected experiments, and a first partner-ready intervention-ranking product. ## Round Snapshot ### A pre-seed round built around prospective biology. The financing advances one investor-underwritable claim: Galen can predict, choose, measure, and learn from real human-cell interventions before results are known. ### At a glance #### stage: Pre-seed diligence path The public workspace gives investors a complete first-pass view before requesting private proof files, round materials, and founder follow-up. #### capital use: 18-month seed proof package Capital funds the human-cell learning loop, model-selected experiments, scientific software, and partner-ready intervention-ranking workflow. #### lead fit: AI-for-science platform conviction Best-fit investors can help with biology leadership, design-partner access, wet-lab judgment, and seed-stage company building. #### approved access: Terms, budget, and exact proof files Approved investors can review use-of-proceeds detail, round materials, baselines, controls, validation contracts, and protocol-level notes. ## Forwardable Intro Email Subject: Galen pre-seed: causal virtual-cell models for intervention prediction Hi [Name], I thought Galen was worth a look for the pre-seed round. Galen is building causal virtual-cell models that predict how living cells respond to intervention, quantify confidence, and choose the next experiment most likely to teach the model. The pre-seed round funds the real human-cell proof loop: predictions committed before results, model-selected experiments, and a first partner-ready intervention-ranking product. The first commercial wedge is intervention ranking for discovery teams deciding which biological changes deserve lab budget. Public investor workspace: https://preseed.usegalen.com Concise partner memo: https://preseed.usegalen.com/partner-memo.md Public proof ledger: https://preseed.usegalen.com/proof-ledger.md AI diligence prompt: https://preseed.usegalen.com/ai-diligence-prompt.txt Private packet request: https://preseed.usegalen.com#data-room The public proof supports controlled causal recovery and disciplined controls. The round funds the prospective human-cell proof loop, model-selected experiments, and a partner-ready intervention-ranking workflow. This is worth a founder conversation for investors who believe intervention prediction is the valuable layer of the virtual-cell stack. ## Five-Minute Diligence Path ### 01: Read the memo Time: 2 min Understand the thesis, the first paid wedge, and why intervention prediction is the valuable layer. Action: Open memo (#memo) ### 02: Inspect the evidence Time: 3 min See what Galen has validated, what is being built, and what the next proof must show. Action: Open ledger (#proof) ### 03: Review the proof plan Time: 3 min Connect pre-seed capital to the forward-looking human-cell study, model-selected experiments, and seed-ready outputs. Action: Open roadmap (#roadmap) ### 04: Request the private packet Time: 1 min Approved investors receive the proof files, round materials, use-of-proceeds detail, and founder follow-up. Action: Request access (#data-room) AI diligence workflow: Open or copy the prompt at /preseed/ai-diligence-prompt.txt to evaluate this public packet with an AI assistant while keeping private proof files gated. Partner forwarding workflow: Open or copy the partner memo at /preseed/partner-memo.md for a concise internal note. Proof workflow: Open or copy the public proof ledger at /preseed/proof-ledger.md to review each claim, current stage, interpretation, and next proof. First-call workflow: Open or copy the diligence questions at /preseed/diligence-questions.md to prepare a founder conversation. Warm-intro workflow: Open or copy the intro email at /preseed/intro-email.txt for a short forwarding note. ## Investor Decision Frame ### Intervention prediction is the valuable virtual-cell layer. Decision: Underwrite Galen if intervention prediction and experiment selection are the valuable layer above biological maps. Public signal: Why-now, product output, and wedge-to-platform sections. Approved packet: Commercial wedge, design-partner workflow, and round materials. ### Causal structure improves prediction before results are known. Decision: Public proof supports a technical call on causal recovery, confidence, and experiment selection for living-cell validation. Public signal: Proof scorecard: causal recovery, controls, experiment selection, and prospective plan. Approved packet: Setup, baselines, ablations, intervals, leakage checks, controls, and exclusions. ### Pre-seed capital buys a seed-ready proof package. Decision: Capital converts modeling evidence into human-cell proof, model-selected experiments, and a partner-ready workflow. Public signal: 18-month roadmap and capital-to-proof plan. Approved packet: Budget, hiring sequence, validation contract, and financing materials. ### The right investor can increase the slope of the proof. Decision: Best-fit investors bring AI-for-science conviction, biology hiring leverage, design-partner access, and seed storytelling. Public signal: Investor-fit section and private-packet preview. Approved packet: Partner pipeline, technical diligence priorities, and round participation. ## Why Now ### Virtual cells are becoming inevitable. Large biological datasets, single-cell atlases, perturbation screens, and foundation-model investment have made the category legible to customers and investors. ### Intervention prediction is the valuable layer. Discovery teams already have maps of biological state. The scarce decision is what to perturb, what response to expect, and which experiment resolves uncertainty fastest. ### The next proof can be scored before results are known. Galen's pre-seed plan converts controlled modeling evidence into a forward-looking human-cell study with predictions, comparison models, and scoring fixed in advance. ### Intervention ranking is useful before the platform is broad. A focused model can help teams prioritize expensive follow-up work while each deployment strengthens the same causal virtual-cell machinery. ## Category Positioning ### maps: Atlases and foundation models make biology legible. State models are the substrate. Galen builds the action layer above them: response prediction, confidence, and experiment choice. Buyer question: What state is this cell in, and what biological programs are visible? Galen position: Galen uses that legibility to predict how the state changes under intervention and how much confidence each recommendation deserves. ### lab: Wet-lab execution turns predictions into measured learning. The lab supplies reality. Galen's product is the model-directed learning loop that decides what to measure and how the result should update the virtual cell. Buyer question: Which experiment should consume scarce lab budget next? Galen position: Galen chooses the experiments with the highest learning value, records predictions before results, and turns each measurement into model improvement. ### workflow: Discovery workflow tools organize decisions. The first product plugs into an existing discovery decision: prioritize interventions before the next expensive experiment queue is run. Buyer question: Which target, perturbation, donor, state, dose, or combination deserves follow-up? Galen position: Galen returns a ranked intervention plan, expected response, confidence, and the smallest useful validation set. ### platform: A focused wedge compounds into programmable biology. The platform is earned through measured performance, reliable confidence, and recurring customer workflows. Buyer question: Can focused prospective reliability expand into a reusable virtual-cell platform? Galen position: Galen starts with causal intervention ranking and expands the same machinery into predict, design, and verify capabilities as evidence accumulates. ## Public Proof Scorecard ### ~90%: Hidden mechanism support recovered in the strongest public summary tests. Galen's current flagship evidence shows the architecture can recover and use cause-and-effect structure when a realistic single-cell-style experiment contains identifiable signal. Private packet note: Exact setup, baselines, confidence intervals, ablations, and leakage checks are in the approved proof packet. ### 15 / 15: Control settings behaved as expected. The evidence practice matters as much as the benchmark: Galen records when evidence supports a claim, when confidence is earned, and when a control outcome remains a control outcome. Private packet note: Approved investors can review the control settings, control outcomes, and claim-boundary notes. ### built: The experiment-selection software is already part of the system. In controlled dry-lab tests, Galen selected experiments expected to teach the model more than simple same-budget alternatives. The pre-seed loop tests that claim in living cells. Private packet note: The packet includes the active-design verdict, comparison policies, and the real-lab scoring plan. ### prospective: The round funds predictions committed before results. The next investor-grade evidence is forward-looking: define the benchmark, lock the baselines, run the experiment, and score the model after biology answers. Private packet note: Approved investors receive the validation contract, use-of-proceeds detail, and protocol-level exclusions. ## Customer Pain ### Every candidate consumes scarce lab budget. Target, intervention, donor, dose, state, and combination choices expand faster than a physical lab can test them with equal discipline. ### Teams need confidence with every recommendation. A useful system must say which prediction is strong, which one is fragile, and which experiment would make the next decision clearer. ### Slow validation weakens the discovery loop. When every decision waits on broad physical search, teams spend months learning which parts of the search space deserved attention. ## Product Output Example ### Suppress target A Predicted response: shift toward desired responder state Confidence: high Next experiment: validate response in focused donor panel ### Add molecule B Predicted response: mixed activation response Confidence: medium Next experiment: measure dose sensitivity before scale-up ### Combine A + B Predicted response: largest expected movement, wider uncertainty Confidence: explore Next experiment: run as learning-rich experiment ## Customer Pull And Commercial Motion ### Teams with expensive intervention choices. Functional-genomics, immunology, target-discovery, and cell-engineering groups already spend budget deciding which targets, perturbations, donors, states, and readouts deserve reality. ### A paid prioritization program with a validation plan. A customer brings a candidate set and a consequential biological decision. Galen returns ranked interventions, expected response, confidence, explicit abstentions, and the smallest useful follow-up experiment set. ### Each deployment improves the same virtual-cell asset. The wedge begins as bounded causal triage. As prospective reliability broadens, Galen can expand into recurring workflow access and programmable predict, design, and verify interfaces. ## Wedge to Platform ### 01: Rank interventions Start with customer decisions that already exist: which targets, perturbations, or follow-up experiments deserve the next tranche of lab budget. ### 02: Commit predictions Record expected response, confidence, comparison models, and scoring rules before the result is known. ### 03: Measure and update Use the lab result to improve the model, the measurement contract, and the next experiment queue. ### 04: Compound the virtual cell Expand reliable prediction across contexts, donors, cell states, interventions, and customer workflows. ## Deck ### Biology has maps. Discovery needs intervention models. Single-cell atlases and foundation models describe biological state. Galen focuses on the next question: what happens when a gene is suppressed, a drug is added, a dose changes, or a cell state shifts. That gap keeps discovery teams in expensive physical search. Every target, intervention, donor, cell state, and combination can demand new laboratory work before a decision can be made. - Which interventions are worth testing? - What response should we expect? - How much should we trust the prediction? - What experiment should be run next? ### Galen is building the virtual cell as the product. The first product is causal intervention ranking and experiment selection. A partner brings a list of possible targets or interventions, relevant assay data, and a consequential decision. Galen returns which interventions are most worth testing, what response to expect, confidence for each prediction, and the smallest useful validation plan. The long-term company is one compounding asset: a causal virtual-cell model that gets better as it selects, runs, and learns from the right experiments. ### The wet lab turns predictions into learning. The virtual cell predicts how a cell should respond when biology is changed and where new evidence will teach the model most. The software chooses the experiment with the highest learning value. The lab turns that choice into a measured result. The result updates, sharpens, and improves the model. The economic goal is more learning per experiment: more decisive measurements, better prioritization, and a model whose reliable region expands with every useful result. ### Current evidence supports causal recovery. Galen's strongest current evidence comes from realistic tests where the true answer was masked from the model. The system inferred which biological changes caused which cell responses and used that knowledge to predict previously unseen responses. This supports a precise claim: Galen's modeling approach can recover and use cause-and-effect structure in controlled settings, and the system is ready for a real lab test scored after predictions are committed. - Beat matched comparison models on previously unseen response prediction in two realistic positive test settings. - Recovered most of the masked cause-and-effect relationships in the strongest public summary tests. - Matched the direction of the masked causal effect in every positive setting. - Correctly separated 15 control settings from signal-bearing settings. ### The first paid use case is target and intervention prioritization. The first customer is a pharmaceutical or biotechnology team in functional genomics, immunology, target discovery, or cell engineering. That team has a biological system, a large candidate set, and a prioritized experimental budget. Galen helps decide which interventions deserve follow-up and which experiments resolve the most consequential uncertainty. ### The pre-seed round funds the decisive conversion step. The pre-seed objective is to convert a controlled modeling proof into a real human-cell lab study, a self-improving learning loop, and first paid partner validation. The round funds the core virtual-cell loop, experimental-biology and lab-automation leadership, modeling and scientific software, measured data from real human-cell experiments, a benchmark scored after predictions are committed, and the first paid partner path. ### If the loop earns trust, the capabilities become a platform. The platform version of Galen exposes simple capabilities: predict how a cell responds when biology is changed, design interventions that move toward a desired state, and verify the experiment that resolves uncertainty fastest. That platform is the long-term reward for making focused virtual-cell models reliable before results are known and useful in customer decisions. ## Proof Summary ### Controlled causal recovery Galen recovered masked cause-and-effect structure in realistic single-cell tests and used it to predict previously unseen responses. ### Experiment selection Galen is measuring whether model-selected experiments teach more than same-budget random or conventional plans. ### Forward-looking human-cell proof The round funds a real human-cell learning loop with predictions committed before results and a second experiment queue chosen from what the model learns. ## Public Proof Ledger ### Causal recovery in realistic tests Status: validated Evidence: In realistic test settings based on measurements from individual cells, Galen recovered the causal relationships behind cell responses, beat matched comparison models in two positive settings, and correctly handled 15 control settings. Interpretation: The model can recover and use cause-and-effect structure when the data contains enough signal to identify it. Current application: Current evidence supports controlled causal recovery under realistic measurement constraints. Next proof: Run the same kind of test on actual intervention data, with predictions and scoring rules fixed before results are known. ### The cause-and-effect path matters Status: validated Evidence: When Galen used its full path for inferring causes, combining effects, and predicting the resulting cell state, it outperformed comparison models built to test whether the signal was real and correctly labeled. Interpretation: The result supports Galen's thesis that causal structure improves intervention prediction. Current application: Current evidence is strongest in controlled settings where the evaluator can score the true causal structure. Next proof: Carry the same discipline into the first real human-cell study. ### Experiment-selection software Status: building Evidence: In a software benchmark, Galen chose a same-budget experiment that taught the model more than random selection. Interpretation: The experiment-selection rule is worth testing in the lab. Current application: This result supports the lab study design for testing model-selected experiments under real experiment cost and noise. Next proof: Compare model-selected experiments against random, diverse, uncertainty-driven, and conventional plans under real experiment cost and noise. ### Integrated virtual-cell machinery Status: building Evidence: The codebase connects cell-response modeling, uncertainty estimates, experiment selection, and reporting into one workflow. Interpretation: Galen is building a coherent system for prediction, experiment choice, measurement, and learning. Current application: The current system is ready to be refined against real human-cell data. Next proof: Use real human-cell data to decide which parts of the system deserve scale-up. ### Real human-cell learning loop Status: next Evidence: The pre-seed roadmap defines a real human-cell study with predictions recorded before results, comparison models fixed in advance, model-selected experiments, and measured outcomes. Interpretation: This is the next value-creating proof. Current application: This study turns Galen's modeling evidence into a living-cell learning loop. Next proof: Run the first forward-looking study in human CD4+ T cells, a central immune-cell type, and score previously unseen interventions before revising the model. ### General programmable virtual-cell platform Status: long-term Evidence: The platform map defines predict, design, and verify as the long-term product layer as focused models become reliable before results are known. Interpretation: The company can become much larger if the loop compounds across contexts. Current application: Galen earns platform breadth by expanding reliable prediction across contexts, interventions, and customer workflows. Next proof: Earn platform language through real performance before results are known, reliable confidence estimates, learning efficiency, and customer workflow integration. ## Pre-seed Roadmap ### Months 0-3: Fundraise story and first validation plan Exit state: Investor evidence pack, first partner workflow, scoring plan fixed before results, candidate experiment queue, and clear data-quality gates. Investor signal: Gives investors a precise story, a concrete first product, and a measurable validation plan. ### Months 3-6: Reliable measurement setup Exit state: Repeatable path from planned cell intervention to measured outcome, quality checks, data provenance, first lab-run plan, and a qualified partner workflow. Investor signal: Creates reliable measurements before the model learns from living biology. ### Months 6-9: First real lab learning cycle Exit state: One cycle where the model predicts, the lab measures, the model updates, and the comparison plan is fixed before results are known. Investor signal: Moves Galen from past-data proof to forward-looking learning. ### Months 9-12: Experiment selection on real data Exit state: Same-budget comparison, second experiment queue, and one paid or formally committed partner validation path if timing permits. Investor signal: Shows whether model-selected experiments produce more learning than volume under real cost and noise. ### Months 12-15: Test across contexts and partners Exit state: Shared intervention data across cell states or donors, partner validation underway, and previously unseen context evaluation if the data supports it. Investor signal: Shows how reusable the learned biology is across cell states, donors, and partner questions. ### Months 15-18: Seed-ready proof package Exit state: Previously unseen result scored after predictions are committed, model-selection result, repeatable partner package, and evidence-based scale plan. Investor signal: Creates the seed-stage claim from measured learning efficiency and partner-ready execution. ## Capital To Proof ### 01: Forward-looking cell proof Run the first real human-cell learning loop with predictions, comparison models, and scoring rules committed before results are known. ### 02: Experiment-selection engine Improve the causal model, confidence estimates, and experiment-selection software around the first validation task. ### 03: Partner-ready package Turn the proof loop into a repeatable intervention-ranking workflow: intake, report, scoring, and next-experiment recommendation. ### 04: Reusable scientific system Build the data provenance, quality standards, and lab-operating practices that make each measured result useful for the next model. ## Founder Notes ### Biology needs models of intervention. The question that matters for discovery is what happens when we act on a cell. #### Maps create the starting point. Biology has become extraordinarily good at mapping. We can measure cells, classify states, observe programs, and build representations over enormous datasets. Galen builds on those maps by modeling how cells respond to intervention. A discovery team eventually has to choose an intervention. It has to decide which target to test, which intervention to advance, which context matters, and which experiment will reduce uncertainty fastest. #### Intervention changes the problem. A useful intervention model handles the moment biology changes: a gene is forced off, a dose changes, a donor shifts, or a combination moves the system into a new state. That is why Galen is built around intervention, calibrated confidence, and testable experiments as the practical layer above representation quality. #### Useful first, broader over time. A virtual cell becomes valuable by making reliable predictions in focused operating domains, assigning confidence, and making the next experiment more decisive. The platform grows by expanding those reliable regions as the model earns trust across contexts. ### The lab is the learning engine. The wet lab is how the model stays answerable to reality. #### The model needs a teacher. A virtual cell becomes useful by staying disciplined by physical biology. The lab supplies the measurements that challenge the model, show where confidence is earned, and decide which predictions deserved trust. The central loop is simple: predict, choose the most informative experiment, measure, update, and turn every result into a sharper specification. #### Control the learning loop. Galen's strategic requirement is control of the learning loop: models, experiment design, protocols, metadata, quality standards, and Galen-funded data. Commodity execution can be purchased or partnered where that is faster and more capital efficient. The moat is model-directed data that is expensive to reproduce economically because each experiment was chosen to expand reliable prediction. #### Every result improves the system. Each experiment scored after predictions are committed updates the model, the measurement contract, or the company thesis. That discipline is how a scientific product earns trust. ### From virtual cell to programmable biology. Predict, design, and verify become a platform as focused models earn trust before results are known. #### The capabilities are simple. A mature virtual-cell platform exposes three capabilities. Predict how a cell state responds to an intervention. Design an intervention that moves the system toward a desired state. Verify the experiment that most efficiently tests the claim. Those capabilities are the interface. The evidence loop beneath them is what makes the interface worth using. #### The first use case compounds. Causal intervention ranking is commercially useful before the platform is broad. Teams already need to rank targets, interpret intervention data, and choose what to test next. Each deployment improves the same virtual-cell machinery. The company becomes a learning system with reusable biology. #### The platform is earned through evidence. Programmable biology is the long arc. Galen earns that position by building reliability across contexts, interventions, and readouts. The near-term job is focused and decisive: prove that a causal virtual cell can make one consequential biological decision more efficient before the result is known. ## Investor Momentum ### Built for teams with expensive prioritization decisions. Functional genomics, immunology, target discovery, and cell-engineering teams already decide which biological changes deserve follow-up. ### Capital converts thesis into measured proof. Capital goes into the first real human-cell learning loop, model-selected experiments, scientific software, and partner validation. ### Seed readiness is evidence-based. The planned output gives Galen a seed-stage story built on forward-looking prediction, measured learning efficiency, and a repeatable partner package. ## Investor Beliefs To Test ### Intervention prediction is the highest-value layer. Maps of biology become more valuable when a team can predict how a cell responds when biology is changed. ### Focused reliability can become a platform. A focused paid use case can improve the same machinery that later supports broader virtual-cell capabilities. ### Galen knows the next proof that matters. The round is organized around a forward-looking human-cell study, calibrated confidence, and model-selected experiments. ### The team can bridge model, biology, and evidence. Galen's founding context combines medicine, machine learning, computational biology, and the discipline to score claims before results are known. ## Data Room Tiers ### Public investor workspace The thesis, first product, proof ledger, roadmap, founder context, FAQ, and public diligence packet. ### Approved investor packet Private proof files, technical model summaries, round materials, use-of-proceeds detail, and founder follow-up. ### Technical diligence Protocol-level details, scoring plans, implementation notes, exclusion rules, and materials that protect Galen's intellectual property. ## Private Packet Index ### Technical proof packet Request: Technical diligence Best for: Investors underwriting model quality, comparison plans, controls, and scientific validity. For a technical read on causal recovery, prediction scoring, confidence, controls, ablations, and the first human-cell validation contract. Includes: - Proof summaries, model-card excerpts, and claim boundaries - Baselines, control outcomes, intervals, leakage checks, and ablation notes - Scoring plan, protocol-level exclusions, and validation contract ### Round materials packet Request: Round materials Best for: Leads and serious followers evaluating ownership, budget, hiring sequence, and seed-readiness plan. For financing diligence on how the pre-seed converts into a prospective proof package, first partner workflow, and seed-stage evidence base. Includes: - Use-of-proceeds detail mapped to proof milestones - Hiring sequence, operating budget, and financing materials - Expected seed-stage evidence package and round follow-up ### Commercial wedge packet Request: Design partner Best for: Investors helping with customer access, buyer workflow diligence, or first design-partner introductions. For a market read on the first paid prioritization workflow, buyer pain, customer intake, and the path from bounded wedge to recurring virtual-cell access. Includes: - First buyer profile and design-partner workflow - Partner intake, report shape, and intervention-ranking deliverable - Commercial motion from first paid program to recurring access ### Founder conversation packet Request: Founder conversation Best for: Investors ready to pressure-test the current proof boundary, next experiment, and founder-market fit. For a focused founder call around technical judgment, scientific prioritization, first hires, and the operating plan for prospective validation. Includes: - Founder-call agenda and highest-signal diligence questions - Current proof boundary and next experiment decision points - Follow-up items matched to the investor's diligence role ## Private Packet Access Path ### 01: Request review We review the request, firm context, and stated diligence need before sharing private proof files or round materials. ### 02: Matched packet Approved investors receive the right package for their intent: proof files, technical diligence, round materials, design-partner detail, or founder follow-up. ### 03: Founder diligence The first conversation focuses on the current proof boundary, the human-cell validation plan, model-selected experiments, and seed-readiness evidence. ## Best-Fit Investors ### AI-for-science and bio-platform conviction. The best-fit investor already believes virtual-cell companies can be large and wants a precise view of Galen's wedge, proof stack, and next falsifiable milestone. ### Capital plus scientific and commercial leverage. The most helpful investors can support senior biology hiring, design-partner introductions, wet-lab execution judgment, and later seed-stage storytelling. ### A seed-ready proof package. The pre-seed round is designed to produce prospective prediction evidence, measured learning efficiency, and a repeatable partner package. ## Private Packet Preview ### Evidence packet Full proof summaries, model-card excerpts, verdict interpretation, baselines, control logic, and claim boundaries. ### Use-of-proceeds and milestone plan Hiring priorities, build sequence, validation contract, capital allocation by proof milestone, and the expected seed-stage evidence package. ### Commercial wedge materials First buyer profile, design-partner workflow, first paid deployment shape, and the path from bounded causal triage to recurring virtual-cell access. ### Founder follow-up A focused diligence conversation around the current proof, the next prospective experiment, and the investor's specific concerns. ## FAQ ### What is a virtual cell? A virtual cell is a computational model intended to predict how a living cell responds when biology is changed, such as suppressing a gene, adding a molecule, shifting state, or testing a combination. ### What is Galen's first product? The first product helps biology teams rank which targets, interventions, or follow-up experiments deserve priority lab budget. ### What has Galen proven? Galen has shown, in controlled realistic tests, that its model can recover masked cause-and-effect structure and use it to predict previously unseen cell responses better than matched comparison models. Galen has also built software that selects experiments expected to teach the model more. ### What is Galen proving next? Galen's next proof is a real human-cell learning loop: predictions recorded before results, model-selected experiments, measured outcomes, and a second round of experiments chosen from what the model learns. ### Why primary CD4+ T cells first? CD4+ T cells are a central human immune-cell type. They connect Galen's modeling work to commercially legible questions in immunology, inflammatory disease, oncology, and cell engineering: which interventions improve function, persistence, activation, or dysfunctional state, and how confident is the model in each recommendation? ### Is Galen a wet-lab company or a software company? Galen is a virtual-cell company. The model is the compounding product, the wet lab is the learning engine, and software is the interface through which predictions, prioritization, and experiment selection become useful. ### What is Galen's differentiated virtual-cell thesis? Galen's thesis is that intervention prediction needs causal structure, uncertainty estimates, and proof before results are known. Large datasets and foundation models validate the category, and Galen differentiates through focused real-world performance, calibrated confidence, learning efficiency, and customer workflow integration. ### What does the pre-seed round fund? The round funds the first real human-cell learning loop, model-selected experiments, scientific software and data infrastructure, and the first paid partner validation.