{ "protocol": "experimental-support-v1", "models": [ "claude-sonnet-5-5", "claude-opus-5-5", "claude-haiku-5-5", "claude-opus-4-8", "claude-fable-5-1", "gpt-6-astra", "gemini-4-argon-high", "grok-4.5" ], "inputSha256": "f1f53a813f414f6a2c30e6b75eaa96e770cf703ad93e32204b69bee8ee561039", "questions": 50, "hypotheses": 250, "policies": [ "advice", "p50", "p70", "p90" ], "pilot": [ { "task": "R42581103", "generator": "gpt-6-astra", "arm": "N", "mapping": [ { "id": "H1", "hypothesisIndex": 3 }, { "id": "H2", "hypothesisIndex": 2 }, { "id": "H3", "hypothesisIndex": 4 }, { "id": "H4", "hypothesisIndex": 0 }, { "id": "H5", "hypothesisIndex": 1 } ], "requests": [ { "protocol": "experimental-support-v1", "system": "You advise researchers which causal hypotheses deserve experimental testing in the biology of ageing.\nUse only the supplied background, research question, hypothesis text and your prior scientific knowledge. You have no tools or internet access. Experimental outcomes are unknown to you.\nForecast experimental support for each hypothesis's own causal claim in its stated biological system. Return a complete JSON object in this reply.", "user": "BACKGROUND\nIntestinal stem-cell function and regenerative capacity decline with age, compromising maintenance and repair of the intestinal epithelium. Both cell-intrinsic changes and signals from the aged systemic environment may contribute to this decline. Chronic inflammation and altered cellular metabolism are common features of ageing, but how circulating conditions are translated into functional defects in intestinal regeneration was unresolved. It was also unclear whether manipulating systemic or metabolic processes could preserve or restore aged stem-cell activity.\n\nRESEARCH QUESTION\nHow does the aged systemic environment causally impair intestinal stem-cell metabolism and regenerative function?\n\nCANDIDATE HYPOTHESES\n[H1] Elevated circulating kynurenine in aged animals persistently activates aryl hydrocarbon receptor signaling in intestinal stem cells, reducing their capacity to use fatty acids for regeneration. Mechanism: Sustained kynurenine–aryl hydrocarbon receptor signaling represses the PPAR-regulated fatty-acid oxidation program, limiting mitochondrial substrate flexibility during regenerative proliferation. Prediction: Kynurenine at concentrations measured in aged plasma reproduces reduced fatty-acid oxidation and regenerative output; stem-cell aryl hydrocarbon receptor deletion prevents both effects despite equivalent kynurenine exposure.\n\n[H2] An age-associated increase in circulating saturated fatty acids impairs intestinal stem cells through ceramide accumulation rather than through inflammatory cytokines alone. Mechanism: Excess saturated fatty acids drive de novo ceramide synthesis in stem cells, disrupting mitochondrial membrane function and increasing reactive oxygen species sufficiently to impair self-renewal. Prediction: The lipid fraction of aged serum reproduces the metabolic and regenerative defects, whereas lipid depletion removes them; inhibiting stem-cell ceramide synthesis rescues function even when aged-serum cytokines remain present.\n\n[H3] The aged systemic environment impairs intestinal stem-cell metabolism primarily by disrupting metabolic support from Paneth cells. Stem-cell defects therefore arise secondarily from an altered local niche. Mechanism: Circulating inflammatory factors suppress Paneth-cell glycolysis and lactate export, reducing the supply of lactate available for oxidation by neighboring intestinal stem cells. Prediction: Paneth cells exposed to aged serum impair otherwise untreated young stem cells after serum washout, with reduced transfer of isotope-labeled lactate into stem-cell mitochondrial metabolites; restoring Paneth-cell lactate export rescues regeneration.\n\n[H4] Elevated circulating TNF in aged animals impairs intestinal regeneration by suppressing mitochondrial pyruvate oxidation in intestinal stem cells. This metabolic restriction directly limits their expansion after injury. Mechanism: TNF activates stem-cell TNFR1–NF-κB signaling, increasing pyruvate dehydrogenase kinase activity and inhibitory phosphorylation of pyruvate dehydrogenase, thereby reducing glucose-derived carbon entry into the tricarboxylic acid cycle. Prediction: Aged serum reduces glucose-derived mitochondrial carbon flux and organoid formation; stem-cell-specific TNFR1 deletion or blockade of pyruvate dehydrogenase inhibition restores both despite continued exposure to aged serum.\n\n[H5] Chronic exposure to circulating interferons in old age depletes intestinal stem-cell NAD+ and thereby compromises mitochondrial function and regeneration. Mechanism: Sustained interferon signaling induces oxidative DNA damage and PARP-dependent NAD+ consumption, reducing NAD+-dependent mitochondrial enzyme activity and respiratory capacity. Prediction: Aged serum increases stem-cell PARylation before NAD+ and respiration decline; interferon-receptor deletion or selective suppression of PARP activity prevents these changes and restores regeneration without correcting circulating lipid or nutrient levels.\n\nAssess EVERY hypothesis independently. Several different mechanisms may be supported, and none may be supported. There is no required number of recommendations and no requirement to recommend one hypothesis per question.\n\nFor each hypothesis estimate p_experimental_support between 0 and 1: your probability that a well-controlled discriminating experiment, competently performed in the species, tissue and scope stated in the hypothesis, would support its central causal claim. Evaluate that claim itself, including a plausible alternative mechanism. This is a subjective scientific forecast, not an observed frequency. If the supplied information cannot justify any numerical estimate, return null and explain the missing information.\n\nGive recommend_for_testing as a separate boolean: would you advise a researcher to spend experimental resources testing this hypothesis now, given its scientific support, testability, assumptions and contradictions? You may recommend multiple hypotheses or none. A recommendation is not a guarantee of success. A numerical estimate is required for a positive recommendation. The comparison will retain your recommendation exactly as issued.\n\nUse the existing scientific considerations: a specific causal chain; consistency with established biology; a feasible discriminating test; fit to the stated question and system; few unsupported assumptions. Novelty is descriptive and is not evidence for or against a claim. Ignore prose style, length, tone, gene-name density and the anonymous candidate order. Do not guess the contents of an undisclosed article or predict another evaluator's score.\n\nProvide a short scientific justification. Quote only literal fragments of the supplied BACKGROUND in support_from_background, or use an empty array. Keep additional assumptions and contradictions separate. Name a concrete intervention, measured readout and necessary controls, and specify outcomes supporting and refuting the claim. These are proposed observations, not results you have seen. Do not announce research, search for information or describe work you will do later; give the complete forecasts now.\n\nReturn exactly this structure, with every supplied candidate once:\n{\"protocol\":\"experimental-support-v1\",\"forecasts\":[{\"id\":\"H1\",\"p_experimental_support\":0.5,\"recommend_for_testing\":false,\"support_from_background\":[],\"assumptions\":[],\"contradictions\":[],\"reason\":\"Short justification\",\"test\":{\"intervention\":\"...\",\"measurement\":\"...\",\"controls\":\"...\",\"supports_if\":\"...\",\"refutes_if\":\"...\"}}]}", "candidateIds": [ "H1", "H2", "H3", "H4", "H5" ], "promptChars": 6963, "promptSha256": "840d59c727e8e8b61b842945ac68ce035b6f698baf5e3161f8298e668f2e4835" } ] }, { "task": "R42601407", "generator": "gpt-6-astra", "arm": "N", "mapping": [ { "id": "H1", "hypothesisIndex": 1 }, { "id": "H2", "hypothesisIndex": 3 }, { "id": "H3", "hypothesisIndex": 4 }, { "id": "H4", "hypothesisIndex": 2 }, { "id": "H5", "hypothesisIndex": 0 } ], "requests": [ { "protocol": "experimental-support-v1", "system": "You advise researchers which causal hypotheses deserve experimental testing in the biology of ageing.\nUse only the supplied background, research question, hypothesis text and your prior scientific knowledge. You have no tools or internet access. Experimental outcomes are unknown to you.\nForecast experimental support for each hypothesis's own causal claim in its stated biological system. Return a complete JSON object in this reply.", "user": "BACKGROUND\nOvarian ageing progressively reduces fertility, disrupts endocrine function and contributes to later-life disease risk. This decline involves losses in follicle quantity and quality as well as deterioration of the ovarian tissue environment. Although lifestyle and systemic metabolic factors have been associated with reproductive ageing, the causal mechanisms connecting them to ovarian function remain poorly understood. No approved treatment currently delays ovarian ageing or reliably extends reproductive lifespan.\n\nRESEARCH QUESTION\nThrough what molecular mechanism does physical activity delay ovarian ageing and extend reproductive lifespan?\n\nCANDIDATE HYPOTHESES\n[H1] Physical activity delays ovarian ageing through muscle-derived irisin, which suppresses profibrotic activation of ovarian stromal fibroblasts. Preserving a compliant extracellular matrix supports follicle survival and sustained reproductive function. Mechanism: Exercise-induced irisin engages αV-containing integrins on ovarian fibroblasts and attenuates TGF-β–SMAD2/3 signalling. Reduced myofibroblast differentiation and collagen deposition limit tissue stiffening and secondary follicular injury. Prediction: Adult muscle-specific Fndc5 deletion should eliminate exercise-associated reductions in ovarian fibrosis and reproductive decline despite comparable training adaptations. Physiological irisin replacement should restore these effects, while fibroblast-restricted disruption of the relevant integrin should prevent rescue.\n\n[H2] Physical activity delays ovarian ageing by repeatedly stimulating endothelial KLF2–eNOS signalling, preserving ovarian microvascular function. Improved perfusion reduces hypoxic injury to growing follicles and maintains steroidogenic capacity. Mechanism: Exercise-associated changes in ovarian vascular shear activate endothelial KLF2 and NOS3, increasing nitric-oxide availability and stabilizing capillary function. Sustained oxygen delivery limits hypoxia-associated oxidative damage and follicular atresia. Prediction: Exercise should improve ovarian perfusion before measurable preservation of follicle survival. Ovarian endothelial Klf2 or Nos3 disruption should prevent these benefits despite intact muscle-derived signals, whereas local restoration of endothelial nitric-oxide signalling should reproduce them in sedentary animals.\n\n[H3] Physical activity delays loss of oocyte quality through circulating apelin acting on granulosa-cell APLNR to enhance mitochondrial quality control. Better granulosa-cell metabolism sustains oocyte developmental competence at older reproductive ages. Mechanism: Exercise-associated apelin activates granulosa-cell AMPK and PINK1–Parkin-dependent mitophagy, removing damaged mitochondria and reducing mitochondrial reactive oxygen species. Improved metabolic support protects the enclosed oocyte from cumulative oxidative injury. Prediction: Adult granulosa-cell Aplnr deletion should abolish exercise-associated increases in mitophagic flux and late-age embryo competence despite preserved vascular and insulin responses. Apelin should reproduce the benefit in sedentary animals, with rescue blocked by granulosa-cell Pink1 deletion.\n\n[H4] Repeated exercise-associated lactate pulses delay ovarian ageing by activating HCAR1 in ovarian macrophages and reducing inflammasome-mediated follicular damage. The benefit depends on transient lactate signalling rather than chronically elevated lactate. Mechanism: Lactate–HCAR1 signalling reduces macrophage NF-κB priming and NLRP3 inflammasome activation, lowering local IL-1β production. Reduced inflammatory signalling limits granulosa-cell injury and follicular atresia. Prediction: Inducible macrophage Hcar1 deletion should abolish exercise-induced reductions in ovarian caspase-1 activation, IL-1β and atresia while preserving systemic metabolic improvements. Lactate pulses matched to exercise exposure should reproduce the ovarian benefit in sedentary controls but fail in the deletion group.\n\n[H5] Moderate physical activity delays ovarian ageing by lowering chronic insulin exposure and preventing excessive primordial-follicle activation. This preserves the follicle reserve through reduced oocyte PI3K–AKT signalling and sustained nuclear FOXO3. Mechanism: Improved skeletal-muscle insulin sensitivity reduces compensatory hyperinsulinaemia, decreasing insulin-receptor signalling in primordial oocytes. Reduced AKT-mediated FOXO3 phosphorylation maintains follicle dormancy and slows reserve depletion. Prediction: In initially hyperinsulinaemic mice, exercise should reduce primordial-oocyte AKT phosphorylation, retain nuclear FOXO3 and extend the age of last successful reproduction. Maintaining sedentary-range insulin exposure during training under matched glycaemia should abolish these benefits, whereas lowering insulin in sedentary animals should reproduce them.\n\nAssess EVERY hypothesis independently. Several different mechanisms may be supported, and none may be supported. There is no required number of recommendations and no requirement to recommend one hypothesis per question.\n\nFor each hypothesis estimate p_experimental_support between 0 and 1: your probability that a well-controlled discriminating experiment, competently performed in the species, tissue and scope stated in the hypothesis, would support its central causal claim. Evaluate that claim itself, including a plausible alternative mechanism. This is a subjective scientific forecast, not an observed frequency. If the supplied information cannot justify any numerical estimate, return null and explain the missing information.\n\nGive recommend_for_testing as a separate boolean: would you advise a researcher to spend experimental resources testing this hypothesis now, given its scientific support, testability, assumptions and contradictions? You may recommend multiple hypotheses or none. A recommendation is not a guarantee of success. A numerical estimate is required for a positive recommendation. The comparison will retain your recommendation exactly as issued.\n\nUse the existing scientific considerations: a specific causal chain; consistency with established biology; a feasible discriminating test; fit to the stated question and system; few unsupported assumptions. Novelty is descriptive and is not evidence for or against a claim. Ignore prose style, length, tone, gene-name density and the anonymous candidate order. Do not guess the contents of an undisclosed article or predict another evaluator's score.\n\nProvide a short scientific justification. Quote only literal fragments of the supplied BACKGROUND in support_from_background, or use an empty array. Keep additional assumptions and contradictions separate. Name a concrete intervention, measured readout and necessary controls, and specify outcomes supporting and refuting the claim. These are proposed observations, not results you have seen. Do not announce research, search for information or describe work you will do later; give the complete forecasts now.\n\nReturn exactly this structure, with every supplied candidate once:\n{\"protocol\":\"experimental-support-v1\",\"forecasts\":[{\"id\":\"H1\",\"p_experimental_support\":0.5,\"recommend_for_testing\":false,\"support_from_background\":[],\"assumptions\":[],\"contradictions\":[],\"reason\":\"Short justification\",\"test\":{\"intervention\":\"...\",\"measurement\":\"...\",\"controls\":\"...\",\"supports_if\":\"...\",\"refutes_if\":\"...\"}}]}", "candidateIds": [ "H1", "H2", "H3", "H4", "H5" ], "promptChars": 7874, "promptSha256": "daa0b7b29419cec2f64d42bdb91e840435b8b41836a8cbd467a730cd05c1d43a" } ] }, { "task": "R42605193", "generator": "gpt-6-astra", "arm": "N", "mapping": [ { "id": "H1", "hypothesisIndex": 4 }, { "id": "H2", "hypothesisIndex": 1 }, { "id": "H3", "hypothesisIndex": 2 }, { "id": "H4", "hypothesisIndex": 3 }, { "id": "H5", "hypothesisIndex": 0 } ], "requests": [ { "protocol": "experimental-support-v1", "system": "You advise researchers which causal hypotheses deserve experimental testing in the biology of ageing.\nUse only the supplied background, research question, hypothesis text and your prior scientific knowledge. You have no tools or internet access. Experimental outcomes are unknown to you.\nForecast experimental support for each hypothesis's own causal claim in its stated biological system. Return a complete JSON object in this reply.", "user": "BACKGROUND\nCellular senescence increases with age and can disrupt tissue function through persistent changes in cell state and signaling. Metabolic dysfunction and chronic inflammation are closely associated with this process and with broader age-related deterioration. Circulating factors can coordinate physiological changes across organs, but it remains unclear which factors actively cause senescence rather than merely correlate with ageing. Identifying the responsible signals and their intracellular mechanisms could reveal targets for extending healthspan.\n\nRESEARCH QUESTION\nWhich circulating regulator causally promotes cellular senescence and organismal ageing, and through what intracellular signaling mechanism does it act?\n\nCANDIDATE HYPOTHESES\n[H1] Elevated lipoprotein-associated C16:0 ceramide acts as a circulating driver of cellular senescence and accelerates organismal functional decline. Mechanism: Delivery of C16:0 ceramide to recipient cells increases mitochondrial ceramide, impairs respiratory function and raises mitochondrial reactive oxygen species. Subsequent DNA damage activates p53–p21 and persistent senescence. Prediction: Lipoproteins enriched with C16:0 ceramide at age-relevant concentrations induce persistent senescence more strongly than otherwise matched particles containing very-long-chain ceramides; recipient-cell mitochondrial ceramide clearance prevents this effect and protects tissue function during chronic exposure.\n\n[H2] Circulating ANGPTL2 drives endothelial senescence and thereby accelerates vascular and systemic ageing. Its effects require sustained inflammatory signaling within endothelial cells. Mechanism: ANGPTL2 engages integrin α5β1 and activates Rac1-dependent oxidative stress and NF-κB signaling. The resulting persistent DNA-damage response activates p53–p21 and reinforces inflammatory secretion. Prediction: Selective depletion of ANGPTL2 from aged plasma reduces its senescence-inducing activity in endothelial cultures, and physiological add-back restores it; endothelial disruption of integrin α5β1 signaling blocks the response.\n\n[H3] Elevated circulating CCL11 directly induces senescence in susceptible neural progenitor cells, contributing to age-associated loss of neurogenesis and cognitive function. Mechanism: CCL11 activates progenitor-cell CCR3, increasing intracellular calcium and NADPH-oxidase-dependent reactive oxygen species. Persistent oxidative DNA damage engages ATM–p53–p21 signaling and stabilizes senescence. Prediction: Exposure to age-relevant CCL11 concentrations produces persistent arrest and multiple senescence features in purified neural progenitors; progenitor-specific CCR3 deletion preserves neurogenesis and cognition during chronic exposure without reducing systemic CCL11.\n\n[H4] Chronically elevated circulating resistin promotes senescence in human vascular and adipose stromal cells, linking metabolic inflammation to tissue ageing. Mechanism: Resistin activates a TLR4-dependent signaling pathway involving MyD88, NF-κB and mitochondrial oxidative stress. Sustained signaling establishes p38–p16–RB-mediated arrest and an inflammatory secretory feedback loop. Prediction: Resistin depletion from aged human plasma reduces persistent senescence in human vascular and adipose tissue models, while physiological add-back restores it; TLR4 deletion prevents the response. Neutralization also improves tissue function in a model expressing human resistin.\n\n[H5] Age-associated elevation of circulating activin A causally promotes senescence in tissue stromal cells and accelerates functional decline. Sustained exposure converts a reversible stress response into stable cell-cycle arrest. Mechanism: Activin A activates ACVR2A/ACVR2B–ACVR1B receptor complexes and SMAD2/3 signaling, inducing p15 and p21. Prolonged signaling establishes persistent arrest and a senescence-associated secretory phenotype. Prediction: Maintaining circulating activin A at concentrations measured in old animals induces persistent senescence and functional impairment in young animals; stromal-cell-specific ACVR1B deletion prevents these effects despite equivalent circulating exposure.\n\nAssess EVERY hypothesis independently. Several different mechanisms may be supported, and none may be supported. There is no required number of recommendations and no requirement to recommend one hypothesis per question.\n\nFor each hypothesis estimate p_experimental_support between 0 and 1: your probability that a well-controlled discriminating experiment, competently performed in the species, tissue and scope stated in the hypothesis, would support its central causal claim. Evaluate that claim itself, including a plausible alternative mechanism. This is a subjective scientific forecast, not an observed frequency. If the supplied information cannot justify any numerical estimate, return null and explain the missing information.\n\nGive recommend_for_testing as a separate boolean: would you advise a researcher to spend experimental resources testing this hypothesis now, given its scientific support, testability, assumptions and contradictions? You may recommend multiple hypotheses or none. A recommendation is not a guarantee of success. A numerical estimate is required for a positive recommendation. The comparison will retain your recommendation exactly as issued.\n\nUse the existing scientific considerations: a specific causal chain; consistency with established biology; a feasible discriminating test; fit to the stated question and system; few unsupported assumptions. Novelty is descriptive and is not evidence for or against a claim. Ignore prose style, length, tone, gene-name density and the anonymous candidate order. Do not guess the contents of an undisclosed article or predict another evaluator's score.\n\nProvide a short scientific justification. Quote only literal fragments of the supplied BACKGROUND in support_from_background, or use an empty array. Keep additional assumptions and contradictions separate. Name a concrete intervention, measured readout and necessary controls, and specify outcomes supporting and refuting the claim. These are proposed observations, not results you have seen. Do not announce research, search for information or describe work you will do later; give the complete forecasts now.\n\nReturn exactly this structure, with every supplied candidate once:\n{\"protocol\":\"experimental-support-v1\",\"forecasts\":[{\"id\":\"H1\",\"p_experimental_support\":0.5,\"recommend_for_testing\":false,\"support_from_background\":[],\"assumptions\":[],\"contradictions\":[],\"reason\":\"Short justification\",\"test\":{\"intervention\":\"...\",\"measurement\":\"...\",\"controls\":\"...\",\"supports_if\":\"...\",\"refutes_if\":\"...\"}}]}", "candidateIds": [ "H1", "H2", "H3", "H4", "H5" ], "promptChars": 7149, "promptSha256": "5f10f7e636d808f88aedbcb11b07407bf9f267c099f2b405126eafbd6f61c3b9" } ] } ], "expectation": "No predetermined winner; higher experimental-support estimates may include alternatives with no matching saved article evidence" }