[ { "id": "EXP-001", "number": 1, "status": "published", "title": "First-look LLM triage: 22 HD abstracts", "question": "What patterns can a local language model identify across recent HD research abstracts?", "method": "Retrieved 22 PubMed papers across five queries and asked Llama 3.1 8B to extract targets, compounds and findings, then propose research hypotheses.", "summary": "The model analyzed 20 of 22 abstracts and surfaced recurring themes in huntingtin modifications, biomarkers and genetic mechanisms. These are exploratory literature signals, not clinical findings.", "limitations": [ "Two abstracts were too short for analysis.", "Abstract-only inputs omit methods and detailed results.", "Model-generated hypotheses require source checking and expert review; this is not medical advice." ], "run_state": "completed", "published_date": "2026-03-27", "model": "llama3.1:8b", "report": "data/experiment_001_report.md", "results": "data/experiment_001_results.json", "page": "experiment-1.html", "reviewed_sources": { "data/experiment_001_report.md": "f799200098a43225c7f279917d44f9b966ec4b3c479ded1ec0d34129cdfa8690", "data/experiment_001_results.json": "2b59a5655628acf150f774e6f85cc92f75177a873baf344613ed856ae669f993" }, "version": 1, "supersedes": null, "superseded_by": null }, { "id": "EXP-002", "number": 2, "status": "published", "title": "Full-text analysis of 16 HD research papers", "question": "What additional research connections emerge when the model reads full-text material instead of abstracts?", "method": "Retrieved 16 open-access papers from PubMed Central, analyzed them with Llama 3.1 8B and synthesized the individual analyses across papers.", "summary": "The synthesis highlighted LIG1 repair fidelity, genetic interventions and mitochondrial-inflammatory connections. Proposed links remain model-generated hypotheses rather than established treatment effects.", "limitations": [ "Only open-access papers were included.", "The 8K context window truncated long papers to approximately 6,000 characters despite the full-text retrieval.", "Single model, single run, without HD domain-expert review; not medical advice." ], "run_state": "completed", "published_date": "2026-03-30", "model": "llama3.1:8b", "report": "data/experiment_002_report.md", "results": "data/experiment_002_results.json", "page": "experiment-2.html", "reviewed_sources": { "data/experiment_002_report.md": "c396b176e5202b45e5406934d7a740031e07d8dce745677d6929d1c1b59869a2", "data/experiment_002_results.json": "02e19f6b36d855b57a7e5918acb3b3bcad1e58740dea1b51a6b8370497302dd8" }, "version": 1, "supersedes": null, "superseded_by": null }, { "id": "EXP-003", "number": 3, "status": "published", "title": "Qwen model comparison: an incomplete run", "question": "How does Qwen 3.5 27B analyze the same 16-paper corpus used in Experiment 2?", "method": "Attempted per-paper analysis and cross-paper synthesis with Qwen 3.5 27B using the Experiment 2 open-access corpus and an 8K context window.", "summary": "The saved results contain inference timeouts, and the report's findings and hypothesis sections are empty. The artifacts document an incomplete comparison, not evidence of model superiority.", "limitations": [ "Inference timeouts prevented a complete analysis and synthesis.", "Paper and character totals describe the input corpus, not successful model comprehension.", "No reliable comparative performance or clinical conclusion can be drawn from this run." ], "run_state": "partial", "published_date": "2026-04-02", "model": "qwen3.5:27b", "report": "data/experiment_003_report.md", "results": "data/experiment_003_results.json", "page": null, "reviewed_sources": { "data/experiment_003_report.md": "651a9a07acc1616f442208ee85d5088f2735b8f2f4a15b4a5a3512be56d77509", "data/experiment_003_results.json": "804f0c3a82bc6a70efffbf6ad223a18511b769f9671dbe3390feb897e1963f28" }, "version": 1, "supersedes": null, "superseded_by": "EXP-003-V2" }, { "id": "EXP-003-GEMMA", "number": 3, "status": "published", "title": "Gemma 4 analysis of the 16-paper corpus", "question": "Which cross-paper mechanisms does Gemma 4 highlight in the Experiment 2 corpus?", "method": "Analyzed the same 16 open-access papers with Gemma 4 and generated a cross-paper synthesis from the individual analyses.", "summary": "Gemma proposed connections among protein aggregation, mitochondrial dysfunction, inflammation and metabolic regulation. This second Experiment 3 publication offers another model's interpretation, not a validated ranking against Llama.", "limitations": [ "The 8K context window truncates long papers; retrieval size is not the amount the model read.", "Cross-disease analogies and model-generated hypotheses are not established HD mechanisms.", "Single run without domain-expert review; not medical advice." ], "run_state": "completed", "published_date": "2026-04-02", "model": "gemma4:latest", "report": "data/experiment_003_gemma4_report.md", "results": "data/experiment_003_gemma4_results.json", "page": "experiment-3.html", "reviewed_sources": { "data/experiment_003_gemma4_report.md": "1cdbe22fc175d44e06569e97bf30c58048b6e07b761604031f62999d75f0352c", "data/experiment_003_gemma4_results.json": "028f7f88a26e8681074eba6989e39a7439cea59822f80dff9d36066586a17880" }, "version": 1, "supersedes": null, "superseded_by": null }, { "id": "EXP-004", "number": 4, "status": "published", "title": "Somatic CAG expansion literature screen", "question": "What research signals concern MSH3, FAN1, PMS1, MLH1 and LIG1 in somatic CAG expansion?", "method": "Searched PubMed with eight targeted queries and screened 75 papers with Gemma 4: 55 full-text inputs and 20 abstracts.", "summary": "The saved artifacts document a targeted DNA-repair literature screen. The report's synthesis and ranking sections are empty, so the screen does not supply a complete published candidate ranking.", "limitations": [ "Top findings, target rankings and hypothesis sections are empty in the report.", "Abstract-only inputs lack full methodological detail.", "AI screening is not medicinal-chemistry review or clinical validation; not medical advice." ], "run_state": "partial", "published_date": "2026-04-07", "model": "gemma4:latest", "report": "data/experiment_004_somatic_cag_report.md", "results": "data/experiment_004_somatic_cag_results.json", "page": null, "reviewed_sources": { "data/experiment_004_somatic_cag_report.md": "0e7ac6bea1b0d1adb2db6887ccfd538d959fb1501cf69ce078cd0ea0deea0a53", "data/experiment_004_somatic_cag_results.json": "c88ece0eb34944397c4fc4cc7f520785a7caebc2e881b58f1c66328684a55245" }, "version": 1, "supersedes": null, "superseded_by": "EXP-004-V2" }, { "id": "EXP-005", "number": 5, "status": "published", "title": "Expanded corpus: genetic and metabolic hypotheses", "question": "Does an expanded corpus reveal research signals beyond somatic CAG expansion?", "method": "After the corpus grew to 65 papers, analyzed 25 papers with Gemma 4, including 16 full-text inputs and nine abstracts, and re-synthesized recurring themes.", "summary": "The report proposes a two-track framing of genetic damage alongside metabolic and immune dysfunction, highlighting copper homeostasis and gut-brain research as topics for further investigation.", "limitations": [ "Only 25 papers were analyzed in this run; the 65-paper corpus total is not the analysis count.", "The two-track framing is an exploratory synthesis, not a validated disease model.", "Candidate mentions and historical trial descriptions are not treatment recommendations or current trial verification." ], "run_state": "completed", "published_date": "2026-04-19", "model": "gemma4:latest", "report": "data/experiment_005_report.md", "results": "data/experiment_005_expanded_results.json", "page": null, "reviewed_sources": { "data/experiment_005_report.md": "0ce8b60dff32cfe4737eb1137eaf70296b2835d9d1169bd4f80469f5a145e878", "data/experiment_005_expanded_results.json": "88f6bf6f78ab34ea9d4e087631b9fa90c4c3b17636f3d84b17bb32ccf9a168c4" }, "version": 1, "supersedes": null, "superseded_by": null }, { "id": "EXP-006", "number": 6, "status": "published", "title": "June 2026 literature update", "question": "Which themes in newly collected HD papers extend the earlier corpus analysis?", "method": "Analyzed 28 papers from the preceding 30-day search window with Gemma 4, using eight full-text inputs and 20 abstracts.", "summary": "The synthesis highlighted membrane lipid interactions, UCHL3-related proteostasis and DNA-repair or epigenetic pathways. These signals motivated subsequent focused literature experiments.", "limitations": [ "Most inputs were abstracts rather than full papers.", "The report's validation language and hypothesis scores reflect model interpretation, not independent experimental verification.", "The June 2026 snapshot is not a current clinical assessment; not medical advice." ], "run_state": "completed", "published_date": "2026-06-27", "model": "gemma4:latest", "report": "data/experiment_006_june2026_report.md", "results": "data/experiment_006_june2026_results.json", "page": null, "reviewed_sources": { "data/experiment_006_june2026_report.md": "afcb5a811c71a576a4ec1571d01b8c4e3b9a1393bccff6f4f9fb9a4fd37ce110", "data/experiment_006_june2026_results.json": "0eaaa63a94fc065b40eff962980fe85a1be2fdcb7248c71585d814da85b3dca7" }, "version": 1, "supersedes": null, "superseded_by": null }, { "id": "EXP-007", "number": 7, "status": "published", "title": "Lipid disruption and protein clearance in HD", "question": "Does lipid raft disruption precede and drive proteostasis failure in HD?", "method": "Used Gemma 4 to analyze 30 papers, split between 15 full-text inputs and 15 abstracts, and synthesize a proposed ordering of mechanisms.", "summary": "The model favored a lipid-first hypothesis connecting membrane organization, mitochondrial stress and protein clearance. The proposed order is a research question, not demonstrated causality.", "limitations": [ "Model confidence scores are not calibrated probabilities or independent verification.", "A literature synthesis cannot establish the timing or causal direction of these mechanisms.", "Proposed intervention targets require experimental testing; not medical advice." ], "run_state": "completed", "published_date": "2026-06-27", "model": "gemma4:latest", "report": "data/experiment_007_lipid_proteostasis_report.md", "results": "data/experiment_007_lipid_proteostasis_results.json", "page": "experiment-7.html", "reviewed_sources": { "data/experiment_007_lipid_proteostasis_report.md": "89e5ad1318cf562e9fd83bd1b2e2249f5a3fa40ff81a810ff5c9d9101d0df399", "data/experiment_007_lipid_proteostasis_results.json": "c8951f95b3f09c84fcf3d464e8a07d966e8180d1fd43b3438b230fdaa479ca26" }, "version": 1, "supersedes": null, "superseded_by": null }, { "id": "EXP-008", "number": 8, "status": "published", "title": "TET1 and epigenetic regulation in HD", "question": "Could TET1 and DNA methylation changes contribute to early HD biology alongside lipid disruption?", "method": "Analyzed 34 papers with Gemma 4, including 18 full-text inputs and 16 abstracts, and synthesized epigenetic and post-translational research signals.", "summary": "The model proposed TET1 suppression as a contributing factor that may interact with lipid and proteostasis pathways. Candidate compounds in the report are hypotheses for research, not evidence of benefit or safety in people.", "limitations": [ "The synthesis does not establish pre-symptomatic causality or an effective intervention window.", "Model-generated candidate scores do not demonstrate clinical safety or efficacy.", "Cross-pathway connections require source review and experimental validation; not medical advice." ], "run_state": "completed", "published_date": "2026-06-29", "model": "gemma4:latest", "report": "data/experiment_008_epigenetic_tet1_report.md", "results": "data/experiment_008_epigenetic_tet1_results.json", "page": "experiment-8.html", "reviewed_sources": { "data/experiment_008_epigenetic_tet1_report.md": "c2eb686f2951fb574ed27830e0c7e0abfad70b349c8094d07737d4e93d46b3aa", "data/experiment_008_epigenetic_tet1_results.json": "9b7e165e5283dd5cb95ccfb06510c7075a097e9353a1ba9126106f6e80a8fc3c" }, "version": 1, "supersedes": null, "superseded_by": null }, { "id": "EXP-009", "number": 9, "status": "published", "title": "Gut-brain and neuronal excitability hypotheses", "question": "What evidence could inform gut-targeted research and striatal neuron excitability in HD?", "method": "Analyzed 35 papers with Gemma 4, using 20 full-text inputs and 15 abstracts, and compared peripheral propagation and neuronal excitability hypotheses.", "summary": "The report proposes links among gut barrier function, systemic metabolism and neuronal pathways. It draws on Parkinson's disease analogies, which do not establish that HD begins in the gut.", "limitations": [ "Cross-disease analogies cannot verify a propagation route or causal mechanism in HD.", "The report's strong-evidence wording and confidence scores are model-generated judgments.", "Gut-targeted candidates and neuronal interventions remain research hypotheses, not medical advice." ], "run_state": "completed", "published_date": "2026-06-29", "model": "gemma4:latest", "report": "data/experiment_009_gut_brain_axis_report.md", "results": "data/experiment_009_gut_brain_axis_results.json", "page": "experiment-9.html", "reviewed_sources": { "data/experiment_009_gut_brain_axis_report.md": "e77361ec130c0380df490fa06f4d40e0354142262a83053dcd4ad0d7bf1dafa6", "data/experiment_009_gut_brain_axis_results.json": "4a90d23136cd46f7baa286bf1dc5946be2a684bf4ffc3c3a5f47d9484041824e" }, "version": 1, "supersedes": null, "superseded_by": null }, { "id": "EXP-010", "number": 10, "status": "published", "title": "NAD+ metabolism and mitochondrial research", "question": "Could DNA-repair stress and NAD+ depletion connect to early mitochondrial dysfunction in HD?", "method": "Analyzed 81 papers with Gemma 4 and generated hypotheses concerning NAD+ precursors, PARP1 activity and SIRT1-related metabolism.", "summary": "The model highlighted NAD+ metabolism as a research direction and proposed precursor, PARP and sirtuin-related hypotheses. The report does not demonstrate that any candidate improves outcomes in HD.", "limitations": [ "Brain exposure, dosing and off-target effects remain unresolved in the report.", "Candidate scores reflect AI assessment rather than validated treatment effects.", "Trial references are a publication-date snapshot, not current status verification or medical advice." ], "run_state": "completed", "published_date": "2026-07-02", "model": "gemma4:latest", "report": "data/experiment_010_nad_metabolic_report.md", "results": "data/experiment_010_results.json", "page": "experiment-10.html", "reviewed_sources": { "data/experiment_010_nad_metabolic_report.md": "6be4e2fbd1e98c782c79b85944893bd428976ee817f473158a0f7b9f3b4b6f86", "data/experiment_010_results.json": "cbf12be3e24e9a2b3dec4fcc70ab2a041aa1e04bd8513fc3a56a1064cf7d9511" }, "version": 1, "supersedes": null, "superseded_by": null }, { "id": "EXP-011", "number": 11, "status": "published", "title": "A footwear concept for chorea-related instability", "question": "How does assumed outsole width change a simplified lateral support model?", "method": "An inspectable 3D design hypothesis with a first-order margin calculation, uncalibrated trade-off sketch and source audit. Original report and drawings are preserved with explicit corrections.", "summary": "The October 7 concept revision starts at an illustrative 115 mm width, with the archived 174 mm width available separately. Broad straps, pull loops, open fastening and heel/forefoot cells remain untested design choices. Neither width is a fit recommendation. The archived 83% is a model ratio, not a measured benefit. No tested shoe or suitable personal dimensions have been established.", "limitations": [ "No prototype, gait-lab validation or human-subject study; not for wear.", "The report cites retracted PMID 24677581. Do not rely on its disease-severity claim or the report's equipment inference. Retraction notice: PMID 28078767.", "Model zero is not a fall threshold. Usable support, stepping, pressure and clearance are not measured; archived dimensions differ between drawings.", "Trade-off optima follow arbitrary coefficients, not patient data. Added distal mass is speculative.", "The cited seven-person HD gait study did not test footwear; base-width variability was nonsignificant (p = 0.064), despite its Discussion's wording.", "Named organizations are capability examples, not partners or endorsements." ], "run_state": "design", "published_date": "2026-07-15", "model": "AI-assisted design study; model not recorded", "report": "data/experiment_011_chorea_footwear_report.md", "results": null, "page": "experiment-11.html", "image": "media/experiment-11/lab/concept-poster-v3.png", "image_alt": "Revised 115 mm illustrative footwear concept with broad straps and pull loops. Dimensions are not a fit recommendation. Unbuilt, unvalidated, not for wear.", "reviewed_sources": { "media/experiment-11/lab/concept-poster-v3.png": "49168eb2a6f02614bdd615dcccdba49477633a8fc153025d636ce85c9f6eb86b", "data/experiment_011_chorea_footwear_report.md": "268165816f44a7ba051c801fedd28f628e30c7ea923065e68d1b8ddfee8431b2" }, "version": 1, "supersedes": null, "superseded_by": null }, { "id": "EXP-003-V2", "number": 3, "status": "published", "title": "Qwen comparison: the saved run timed out", "question": "Did the April 2 Qwen 3.5 run produce findings that can be compared with Llama?", "method": "Counted analysis errors and the synthesis object in experiment_003_results.json. Ollama was not called again.", "summary": "All 16 paper analyses and the cross-paper synthesis failed with an Ollama read timeout at localhost:11434. The file records input size only. No Qwen finding was produced.", "limitations": [ "This closes an empty report; it is not a model comparison.", "Titles were not stored, so none are reconstructed.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "qwen3.5:27b (artifact audit; model not rerun)", "report": "data/experiment_003_v2_report.md", "results": "data/experiment_003_results.json", "page": null, "version": 2, "supersedes": "EXP-003", "superseded_by": null, "reviewed_sources": { "data/experiment_003_v2_report.md": "d44ee13546320da20dcbd102fbba585c33605e5839ab4fbfea17c666721a9467", "data/experiment_003_results.json": "804f0c3a82bc6a70efffbf6ad223a18511b769f9671dbe3390feb897e1963f28" } }, { "id": "EXP-004-V2", "number": 4, "status": "published", "title": "Somatic CAG screen transcribed from the saved Gemma synthesis", "question": "What target and modality rankings did the April 7 Gemma run actually save?", "method": "Transcribed the synthesis object from experiment_004_somatic_cag_results.json. No model was rerun.", "summary": "The saved synthesis ranks MSH3, PMS1 and FAN1 and records 68 parsed analyses plus 7 JSON parse failures. Those ranks are the model's assessment, not a new measurement.", "limitations": [ "Seven paper analyses failed to parse and are excluded from the synthesis.", "Ranks are not clinical effect sizes.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "gemma4:latest (artifact audit; model not rerun)", "report": "data/experiment_004_v2_report.md", "results": "data/experiment_004_somatic_cag_results.json", "page": "experiment-4.html", "version": 2, "supersedes": "EXP-004", "superseded_by": null, "reviewed_sources": { "data/experiment_004_v2_report.md": "9005e9ebe11ffb6b7f3ede57824517ae4c39d906513685135e2dbaf2072feffa", "data/experiment_004_somatic_cag_results.json": "c88ece0eb34944397c4fc4cc7f520785a7caebc2e881b58f1c66328684a55245" } }, { "id": "EXP-024", "number": 24, "status": "published", "title": "HDinHD access probe: no public catalog", "question": "Can this hub ingest HDinHD perturbation, omics, gene-set, animal-model, and GeM-HD data without an account?", "method": "HTTP probe of the description page, the account form, and explorer.hdinhd.org, with redirects followed and no login.", "summary": "Explorer redirects to a login page. The probe saved zero catalog rows. A normalizer and a synthetic fixture exist for when HDINHD_TOKEN is available.", "limitations": [ "Logged-out responses cannot prove what an approved account can export.", "No HDinHD study counts were copied from the marketing page.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "none (HTTP probe)", "report": "data/experiment_024_hdinhd_report.md", "results": "data/hdinhd_access.json", "page": "experiment-24.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_024_hdinhd_report.md": "cda9b6e6a4b64fb49436ac8bd46a6e2db603a49c5a85b43410bccfcf468b5928", "data/hdinhd_access.json": "c1965c4d00274bde0f5e16f2e1d6242f3a488c6408364139fe1b51bcab1fad5a" } }, { "id": "EXP-025", "number": 25, "status": "published", "title": "Paper corpus catch-up after the May 5 stall", "question": "Did papers stop updating because PubMed was empty, or because paper_scout was not scheduled?", "method": "Compared the May 5 corpus and pipeline log with the daily workflow, then ran paper_scout --discover-only. A 429 on the first detail fetch was retried.", "summary": "The scout corpus had stopped on 2026-05-05 at 86 papers while data_fetcher kept refreshing homepage papers. Discover-only search raised the corpus to 156 papers and analyzed none, because no LLM is configured.", "limitations": [ "Broad queries saved some non-HD titles; they were not rescored.", "NCBI_API_KEY was not set.", "Scripts 12-23 were not executed; their published accounts only record that block.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "none (PubMed discovery)", "report": "data/experiment_025_paper_refresh_report.md", "results": "data/experiment_025_paper_refresh_results.json", "page": "experiment-25.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_025_paper_refresh_report.md": "e3fb7c6dd8118dd8de71cc860ac4f4078083058e0dcdcd20658e0e34e76e9b86", "data/experiment_025_paper_refresh_results.json": "5d065b3e4dd64407c914f71053b1fa32f0c11c342eb30de03c289525e1e22814" } }, { "id": "EXP-012", "number": 12, "status": "published", "title": "Kynurenine Pathway Dysregulation in HD \u2014 KMO, IDO1, and QUIN/KYNA Imbalance", "question": "1. How well does QUIN/KYNA imbalance explain MSN/SPN vulnerability in HD striatum? 2. Can KMO inhibition (CHDI-00340246, Ro61-8048) reduce HD pathology independently of mHTT lowering? 3. Is IDO1 inhibition (epacadostat repurposing) tractable for HD neuropsychiatric symptoms? 4. Does kynurenine pathway dysregulation interact with other HD mechanisms (NAD+/h021, neuroinflam/h001)?", "method": "Checked run_experiment_12_kynurenine_pathway.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_012_blocked_report.md", "results": "data/experiment_012_blocked_results.json", "page": "experiment-12.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_012_blocked_report.md": "47b136d942b5847dd8b21be0b372f9b694e2352768a47af97d7ad84786b537db", "data/experiment_012_blocked_results.json": "f0bc61a751e566458694d5af3797bfbc0505bf8184c3c927db7ca5b4d8ac52c0" } }, { "id": "EXP-013", "number": 13, "status": "published", "title": "Complement Cascade & Neuroinflammation in HD", "question": "Is complement-mediated synaptic pruning a tractable therapeutic target in Huntington's Disease?", "method": "Checked run_experiment_13_complement_neuroinflammation.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_013_blocked_report.md", "results": "data/experiment_013_blocked_results.json", "page": "experiment-13.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_013_blocked_report.md": "3559231821c791031575dd8c7b5010c0c6993daa940539aa050d546ba8822593", "data/experiment_013_blocked_results.json": "856a970a4db16102e864456450c316ae657e207244292dc8a81a0720414c952a" } }, { "id": "EXP-014", "number": 14, "status": "published", "title": "Cognitive Pharmacotherapy in Huntington's Disease", "question": "Which pharmacological approaches best target cognitive decline in HD, and does the cholinergic system offer a tractable repurposing opportunity?", "method": "Checked run_experiment_14_cognitive_pharmacotherapy.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_014_blocked_report.md", "results": "data/experiment_014_blocked_results.json", "page": "experiment-14.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_014_blocked_report.md": "d41903dce1e12f51dd1732ca72cde19f7b6a9354dc6c4341817e38bdbbc1d716", "data/experiment_014_blocked_results.json": "f2619e8e24d7de26b92359935c971ad26f45716cd65b54395d25cdb627842c5d" } }, { "id": "EXP-015", "number": 15, "status": "published", "title": "SEMA4D Neuroinflammation Axis in Huntington's Disease", "question": "Does semaphorin 4D (SEMA4D) / Plexin-B1 signaling represent a tractable upstream neuroinflammation target in HD, and what is the evidence for pepinemab (VX15/2503) beyond the SIGNAL-HD Phase 2 trial?", "method": "Checked run_experiment_15_sema4d_neuroinflammation.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_015_blocked_report.md", "results": "data/experiment_015_blocked_results.json", "page": "experiment-15.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_015_blocked_report.md": "583a9807740234b17a8749a11413a2e8125f3a23f225ac4023fbc9d50cab3e17", "data/experiment_015_blocked_results.json": "56dc743e955279e1d26462180b71eddf5b3241d05a90c6a607a4af0866b87839" } }, { "id": "EXP-016", "number": 16, "status": "published", "title": "Mitochondrial Membrane Stabilization in HD", "question": "1. Is Complex II (succinate dehydrogenase) loss a driver or consequence of mHTT toxicity? 2. Does cardiolipin disruption precede or follow mHTT aggregation in HD neurons? 3. What is the therapeutic window for elamipretide (SS-31) in HD striatum? 4. How does inner mitochondrial membrane (IMM) stabilization compare to NAD+ restoration (h021)? 5. Can combining IMM stabilization (elamipretide) with NAD+ precursors (NR/NMN) rescue energy failure?", "method": "Checked run_experiment_16_mitochondrial_membrane.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_016_blocked_report.md", "results": "data/experiment_016_blocked_results.json", "page": "experiment-16.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_016_blocked_report.md": "1bb92120553ea300fb7b50f3e46ece738d1780f529a5c62ea9c635d39937363e", "data/experiment_016_blocked_results.json": "3a0e8aab56284b08180bd3c82382fdafd14ab42831fdc88cccc61aba6b0d6b01" } }, { "id": "EXP-017", "number": 17, "status": "published", "title": "PDE10A Inhibitors \u2014 Striatal cAMP/cGMP Rescue in HD", "question": "1. Is PDE10A loss in HD striatum a driver of MSN dysfunction or a downstream consequence? 2. Do PDE10A inhibitors restore DARPP-32/CREB/PGC-1\u03b1 signaling in HD preclinical models? 3. What is the presymptomatic therapeutic window for PDE10A targeting in HD? 4. How does the cAMP/cGMP axis interact with existing hypotheses (h020 MSN excitability, h021 NAD+/SIRT1, h037 BDNF/TrkB)? 5. Can PDE10A PET imaging serve as a progression biomarker for HD clinical trials?", "method": "Checked run_experiment_17_pde10a_striatal_rescue.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_017_blocked_report.md", "results": "data/experiment_017_blocked_results.json", "page": "experiment-17.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_017_blocked_report.md": "748bf58c47b758ef166182e83666c6b65a55a7011a0e4be372e30dd47f9bc0c4", "data/experiment_017_blocked_results.json": "e9f6bdc35169d540c0a01f493eedb82c46369797784acd6986cf5251d4222e30" } }, { "id": "EXP-018", "number": 18, "status": "published", "title": "GLP-1R Agonists \u2014 Neuroprotection Repurposing for HD", "question": "1. What is the evidence for GLP-1R expression and function in HD-vulnerable circuits (striatum, cortex, hippocampus)? 2. Do GLP-1R agonists (exendin-4, liraglutide, semaglutide) reduce neurodegeneration in HD mouse models \u2014 and via which pathways? 3. How does the SPARK trial (lixisenatide in PD, Lancet 2024) de-risk the HD repurposing", "method": "Checked run_experiment_18_glp1r_neurodegeneration.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_018_blocked_report.md", "results": "data/experiment_018_blocked_results.json", "page": "experiment-18.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_018_blocked_report.md": "bfd71b9f064cc977ff9494d155cd90c399ab263e8e2c81dc80c502e2a87a11d3", "data/experiment_018_blocked_results.json": "ee5c0f870bb787fb307722d3caf95aa44655ec4a4683ce6a2db11cb95fb980a4" } }, { "id": "EXP-019", "number": 19, "status": "published", "title": "MPC Inhibition & ISR Normalization in HD", "question": "Does inhibiting the Mitochondrial Pyruvate Carrier (MPC) to attenuate Integrated Stress Response (ISR) overactivation represent a viable disease-modification strategy for Huntington's Disease?", "method": "Checked run_experiment_19_mpc_isr_inhibition.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_019_blocked_report.md", "results": "data/experiment_019_blocked_results.json", "page": "experiment-19.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_019_blocked_report.md": "df951e058ee691499e4b8c830402472ed0aa595e49b789f49727006fb3ccd2cc", "data/experiment_019_blocked_results.json": "46ab63e275b7000c00f0de938440e2102dbfe15c3aeb1ea3f50865566049deff" } }, { "id": "EXP-020", "number": 20, "status": "published", "title": "Ferroptosis & Iron Chelation in HD", "question": "Is ferroptosis \u2014 iron-dependent, GPX4-regulated lipid peroxidation cell death \u2014 a primary neurodegeneration mechanism in HD striatal neurons, and can ferroptosis inhibitors or iron chelators slow HD neurodegeneration?", "method": "Checked run_experiment_20_ferroptosis_iron_chelation.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_020_blocked_report.md", "results": "data/experiment_020_blocked_results.json", "page": "experiment-20.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_020_blocked_report.md": "a49da5ea33abc0fe231383420841f6355e5050ef16ec9a8f6b8e2e6c1da0af5a", "data/experiment_020_blocked_results.json": "0648d1d6e6f5e54cc3441cc20223d8d3306866bc4c53b7bc0a3ba4f89cd1ffd1" } }, { "id": "EXP-021", "number": 21, "status": "published", "title": "cGAS-STING Innate Immune Pathway in HD", "question": "Does mHTT-driven mitochondrial DNA (mtDNA) release activate the cGAS-STING innate immune axis in HD striatal neurons, and can STING inhibitors suppress the resulting type I interferon neuroinflammatory cascade to slow HD neurodegeneration?", "method": "Checked run_experiment_21_cgas_sting_innate_immunity.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_021_blocked_report.md", "results": "data/experiment_021_blocked_results.json", "page": "experiment-21.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_021_blocked_report.md": "e0ca2e6da357bf5d9a64a0c69bc85a502a583e0dc27b91e5f2d61e4deceda69d", "data/experiment_021_blocked_results.json": "143f2ffcc94b66899a2eb21bb50c9a541bc7ad14641d3bfdd2bc14dfb4a7118a" } }, { "id": "EXP-022", "number": 22, "status": "published", "title": "Adenosine A2A Receptor Antagonism for HD MSN Protection", "question": "1. What is the mechanistic basis for A2A receptor overactivation in HD striatal MSNs? 2. Does istradefylline rescue HD phenotypes in R6/2 or zQ175 models? 3. What is the A2AR-cAMP-PKA-BDNF signaling cascade in MSN vulnerability? 4. How does A2A antagonism compare with or complement PDE10A inhibition? 5. Is there epidemiological evidence linking caffeine intake to delayed HD onset? 6. What is the druggability and CNS pharmacology of A2A antagonists? 7. Are there HD-specific A2A receptor expressio", "method": "Checked run_experiment_22_adenosine_a2a_msnprotection.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_022_blocked_report.md", "results": "data/experiment_022_blocked_results.json", "page": "experiment-22.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_022_blocked_report.md": "886fd850e27a70423e6ff3b0885ce2f15242afafbeece5defb1a5912333feb05", "data/experiment_022_blocked_results.json": "4ed87ec95fcea8110525453942bcc3a7d092625ed341570ed788ccef939116e8" } }, { "id": "EXP-023", "number": 23, "status": "published", "title": "Thiamine + Biotin TCA Cycle Rescue in HD", "question": "Can high-dose thiamine and biotin restore TCA cycle enzyme activity in HD striatum, and what is the mechanistic overlap with BTBGD?", "method": "Checked run_experiment_23_thiamine_biotin_tca_rescue.py and did not execute it. No LLM backend was reachable.", "summary": "The script was not run. No papers were fetched and no model output was produced. This page is the finished account of that block.", "limitations": [ "No PubMed search was run for this script.", "No model output.", "Not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "not run", "report": "data/experiment_023_blocked_report.md", "results": "data/experiment_023_blocked_results.json", "page": "experiment-23.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_023_blocked_report.md": "b7392f756f40bcf8c4065ab354627c703e0c9f1ef0bdca71b0e6754d46d5f1b2", "data/experiment_023_blocked_results.json": "04e4048c7e9f5039e4c65c88dc000ed33b20f9f531497d403da08dcb86282ace" } }, { "id": "EXP-026", "number": 26, "status": "published", "title": "What the shoe-building note actually connects to", "question": "Is there a shoe manufacturing project, or only the unbuilt EXP-011 design study?", "method": "Searched this repo and jravinder GitHub code, read the EXP-011 audit, and fetched the Formlabs page that audit cites.", "summary": "The only HD shoe work found is Experiment #11, an unbuilt design study merged in PR 282. Named factories are capability examples, not partners. Other shoe hits in Ravi's repos are unrelated.", "limitations": [ "GitHub code search can miss unindexed private text.", "New Balance, Formlabs, Adidas, and Materialise were not contacted.", "Not a product recommendation and not medical advice." ], "run_state": "completed", "published_date": "2026-10-04", "model": "none (repository search)", "report": "data/experiment_026_shoe_tiein_report.md", "results": "data/experiment_026_shoe_tiein_results.json", "page": "experiment-26.html", "version": 1, "supersedes": null, "superseded_by": null, "reviewed_sources": { "data/experiment_026_shoe_tiein_report.md": "00dd247f3b7366680cf7af8611a188e10b6f3f239e1dad2753c393d6002561dd", "data/experiment_026_shoe_tiein_results.json": "4ba95f082593b1f63faacafe95b437ce6f88538a2e05cb3bdf6ab5dd94b01e5c" } } ]