An emerging field
Astro-Genetics
Stargazers and Trailblazers, the IAMPerformance Inter-Domain Research Institute are pioneers in the emerging field of Astro-Genetics. They apply the same rigorous discipline of Cosmic-Cartography to the human epigenome with the analogies (CMB anisotropies → architectural drift signatures; inhomogeneous decoherence at the cosmic horizon → inhomogeneous floor crossings at the epigenome) drawn explicitly.
What EDEAR is, what it isn't
EDEAR gives a cellular operating score. It does not diagnose disease. Qualified medical professionals will determine the appropriate response.
IAMPerformance Inter-Domain Research Institute
I den frie forskers tradition — i taknemmelighed for dem, der gik forud.
In the tradition of the independent researcher — in gratitude to those who came before.

ONE LAW
UNIVERSALLY From the Telescope to the Microscope — Cosmic Scale to Cancer Cell.

Every irreversible transition — cosmic structure formation, qubit decoherence, transistor switching, cellular epigenomics — leaves the same dimensionless fingerprint. One law. Every domain.

Entropy is the same whether it’s driving cosmic expansion or a malignant cell.
Not a metaphor. The same physics measures a cell’s departure from healthy architecture and places a gravitational system on the same thermodynamic scale. IAM is the framework under which both measurements are made, from the same dimensionless fingerprint, across 37 orders of magnitude. Commercial applications are delivered through the APE family of instruments.
Malignancy
A cell that has exceeded its thermodynamic maintenance budget. The A-score crosses the architecture floor. The cell can no longer pay the cost of remaining itself.
Black Hole
A star that did the same. The gravitational encoding surface reaches its physical capacity. The star can no longer pay the cost of remaining itself.
Same physics. Same threshold. 37 orders of magnitude apart.
For technical readers arriving from LinkedIn, GitHub, or a PDF issue
This site shows the same thermodynamic architecture appearing in multiple substrates. The cosmology work is the research foundation. Astro-QAPE, Astro-SCAPE, and Astro-GAPE are the domain-specific commercial instruments. The fastest way to read the site is: common thread first, then the APE issues and demos.
Latest Research · May 2026
Pre-clinical · Not peer-reviewed

The Physics of Pathology
Informational fidelity, thermal noise, and cellular computation.

Every clinical instrument in current use is a form of steam detection — the water has to be boiling for the detector to fire. EDEAR is a thermometer. It reads where each architectural cell class is currently sitting on the temperature scale, calibrated against thermodynamic floors derived from physics, and it reads in both directions. A client running hot has water rising toward a boil. A client running cold has water that has dropped below the healthy operating range. Two readings six months apart give a trajectory. Not a single coordinate.

Headline Result
In samples drawn 10+ years before clinical breast cancer diagnosis, the Stage 1 immune-class A-score reads d = +1.78 (GSE51057, n = 11) and d = +1.36 (GSE51032, n = 33) — the same monotonic trajectory across two independent EPIC-Italy cohorts. Permutation p < 0.0001 (10,000 label shuffles, no exceedances). 50% of future-cancer cases at >10 yr pre-diagnosis sit at or above the 90th percentile of age-matched cancer-free controls.
17
clinical applications
128+
sealed VAL studies
8
architectural classes
0
free parameters
Download Paper (PDF) GitHub — All Scripts Zenodo DOI →
Mahaffey HW (2026). The Physics of Pathology: Informational Fidelity, Thermal Noise, and Cellular Computation. Standalone Technical Evidence Report. IAMPerformance Inter-Domain Research Institute, Entiat WA. Patents pending 64/012,720 & 64/014,568 · Pre-clinical research. Not peer-reviewed.
37
Orders of magnitude
Same law from cosmic horizon to cell nucleus
29/30
Cancer types confirmed
Zero free parameters · Zero training data
0.07%
Cosmological constant
Agreement with observation · Zero free parameters
Astro-GAPE — Astro-Genetic Architecture-class Performance Engine

Cellular operating scores.
Physics, not pattern-matching.

The cellular application of IAM produces a physics-derived index — the A-score — that quantifies how far a cell has departed from its thermodynamic architecture floor. Validated against 4,304 published matched tumor-normal pairs across 30 TCGA cancer types at zero free parameters. Pre-clinical results only. Prospective clinical validation has not been performed.

A = H(β) / H_min(class)
One equation. A mean methylation beta from the cell architecture class of concern — divided by the physics-derived H_min floor for that class — gives the A-score. The threshold A > 1.05 was derived from healthy-cell H_min calibration — not from cancer training data. In published stage-stratified methylation data, this threshold corresponds to the pre-invasive stage in the majority of tested cancer types. What assay most reliably delivers that beta from a blood draw is an open question for prospective study. Pre-clinical research only. Not intended to diagnose, treat, or screen any individual. Prospective clinical validation has not been performed.
Coming Soon
EDEAR — Early Detection · End-of-life · Aging · Research
The clinical product line being built on Astro-GAPE.
EDEAR is the clinical product line currently in development on top of the Astro-GAPE engine. The design intent is straightforward: one Illumina EPIC blood draw — the same standard 850K methylation array every commercial methylation lab already runs — routes through the universal three-stage pipeline (immune-class A-score, tissue-of-origin deconvolution, immune sub-composition) and produces a complete cellular-state report covering seventeen clinical applications in a single workflow: pre-diagnostic cancer signatures, neurodegeneration patterns, hematologic malignancy signatures, cardiovascular trajectories, and inflammatory bowel disease signatures. The framework reads where each architectural cell class is currently sitting on the temperature scale, calibrated against thermodynamic floors derived from physics — not fitted to clinical outcomes. A client at the floor is at the healthy operating temperature for her cell class. A client running hot has water rising toward a boil. The thermometer reads both directions.
17
Clinical cards built
128+
Sealed validation studies
1
Blood draw
📄 EDEAR Operations Summary Pre-clinical research only · Prospective clinical validation has not been performed · Not intended to diagnose any individual
29/30
Cancer types confirmed
4,304 matched pairs
A>1.05
Physics-derived threshold
Not from cancer training data
0.3%
Normal f_C3
Cancer: 13.0%
0
Free parameters
Zero training data used
🩸 The Plasma Problem — Solved via Deconvolution
Bulk plasma cfDNA is dominated by hematopoietic cells (~70%). Tumor-tissue signal is diluted below what architectural scoring can resolve directly. VAL-038 confirmed the framework's own prior prediction: bulk plasma pan-cancer correlation ρ = −0.02 — an honest null. VAL-041 then closed the loop: when plasma is deconvolved per Moss 2018 tissue markers and each per-tissue fraction scored against its class-specific floor, tissue-of-origin localization is correct in 10 of 10 tested cancer types (mean ΔA = +0.174 at the correct tissue). The clinical workflow is plasma draw → tissue-of-origin deconvolution → per-tissue A-score. Not a new assay — a new interpretation layer on existing methylation panels.
🧠 Terminal-Class Cancers — The Framework's Structural Prediction
Neurons carry the tightest architectural floor of any somatic cell class. IAM predicts: the more constrained the healthy starting point, the larger the departure when the architecture fails. Glioma sits at ΔA = +0.306 from the Moss 2018 cfDNA atlas — the largest signal across 30 TCGA cancer types. LGG and GBM rank first and second from first principles, with no cancer training data involved in the calibration. VAL-044 confirms the Stupp protocol (TMZ + RT) responder/non-responder separation; CSF sampling gives ~10× cleaner terminal-class signal than plasma. Brain tumors are the framework's sharpest structural prediction on record.
🔬 Dense Breast — A Physics-Blind Screening Modality
Approximately 40% of women carry dense breast tissue; mammogram sensitivity falls to ~55% in that population (versus ~85% in non-dense). Astro-GAPE measures DNA methylation entropy, not X-ray absorption. The secretory-class floor is an epigenomic property, not an anatomical one — the same for both tissue densities. VAL-047 (EPIC-Italy GSE51057, 13.9-yr follow-up) detected a breast architectural signature a decade before clinical diagnosis using the same buffy-coat DNA a mammography patient could provide on the same day. The question is not whether the physics sees density — it is whether a decade of blood-based runway closes the gap the mammogram leaves open.
📊 DCIS Stratification — The Active vs Inert Question
Over 50,000 US women are diagnosed with DCIS each year; 15–50% may never progress to invasive disease. The clinical problem is stratification, not detection. Published stage-stratified methylation places low-grade DCIS below the A-score threshold and high-grade DCIS above it. Framed against VAL-042's monotonic pre-cancer progression across five organ systems (cervical, Barrett's esophagus, prostate PIN → metastatic, colon adenoma → CRC, and CHIP → MDS → AML — 5/5 monotonic, 4/5 reach floor breach), DCIS fits the same thermodynamic tier structure every other pre-cancer system in the cascade follows. Prospective DCIS stratification study is an open path; the stage-wise A-score trajectory is already visible in retrospective data.
🧬 Alzheimer's — Not a Neurodegeneration, a Multi-Class Architectural Failure
VAL-040 tested whether AD is confined to terminal-class (neuronal) drift or manifests across multiple architecture classes. Result: four of eight classes show elevation in AD cohorts — terminal (brain cortex, expected), immune (peripheral blood, novel), secretory (pancreatic islet via the T2D–AD comorbidity axis), and stromal (cerebral vasculature). Seven of seven tissue-class combinations show a severity gradient from early- to late-stage AD. AD is a systemic multi-class thermodynamic phenomenon, detectable peripherally, across four large independent cohorts: De Jager ROSMAP n=708, Shireby BDR n=1,408, Nabais peripheral-blood meta-analysis n=3,424, Lunnon entorhinal cortex.
The same multi-class drift signature that precedes cancer by 2–5 years appears for AD in independent cohorts. The detection threshold was not trained on AD patients — it was derived from what healthy cells require. If drift precedes clinical AD the way it precedes clinical cancer, the peripheral-blood signal is the opening, not the terminal tissue.
🐕 Cross-Species Confirmation — The Physics Transfers
VAL-043 tested the class-floor framework across five canine cancers (mammary, DLBCL, osteosarcoma, bladder TCC, lymphoma) against the same architecture classes calibrated on human tissue. Mean cross-species A-score difference = 0.010 across 70 million years of evolutionary divergence. Canine aging trajectory r = 0.9995. If the floor is a physics quantity, it should hold across species — and it does. The next step is a prospective Astro-GAPE validation study in dogs with compressed lifespan and cohort-scale tractability: a 12–18 month timeline to independent evidence in a non-human mammal, using the exact same instrument.
Download Astro-GAPE Issue 002 Open Demo 📧 Evidence Report — email author Published research is open · Commercial applications and calibration constants are proprietary
What Astro-GAPE Can Do — Validation Record as of May 2026

Imagine a pot of water on a stove. The burner is set somewhere on the dial — low, medium-low, medium, medium-high, high — now imagine you set the dial to whatever temperature you choose. That determines the temperature your water is going to be. If the burner stays too high for too long, the water reaches a boil. No one is around to see it starts spilling over the rim. Left that way unattended, the water boils away entirely. Nothing you can do now will ever get that water back into the pot. That pot is a human cell population. The burner is whatever has been driving that cell population to do more cellular work than it should — chronic inflammation, chronic infection, carcinogen exposure, autoimmune dysregulation, prolonged hormonal exposure. The water reaching a boil is the moment a tumor crystallizes and starts producing the symptoms that send a patient to a clinic. Unfortunately today, the only thing that can be detected is steam — the water must be at a raging boil to be detected by anything that currently exists. This framework changes that. It reads both the temperature of the water and the setting of the dial. It reads where the cell population is currently sitting on the temperature scale, calibrated against thermodynamic floors derived from physics, and it reads in both directions.

The cards below highlight the most clinically relevant findings from 128+ sealed validation studies (VAL-001 through VAL-128) — spanning the pan-cancer methylation core, the multimodal five-substrate framework, the vertebrate lifespan extension, the multi-class drift cascade, external per-patient 450K/850K replication, and the multi-atlas calibration discipline wave that anchors every active disease card. The cascade tests alone (VAL-037 to VAL-046) confirmed 35 of 39 pre-specified predictions on first attempt. Every result is computed from published primary-source data with zero free parameters and zero cancer training data. Pre-clinical research only; prospective clinical validation required before any individual-patient use.

📅
10+ Year Pre-Diagnostic Window — Breast Cancer
Pink ribbon · Secretory class · VAL-047 · EPIC-Italy GSE51057
The EPIC-Italy cohort is a real-world prospective study: women donated blood at study enrollment with no idea whether they would later develop cancer. Years later, some did. When the framework reads the blood drawn more than ten years before those women would walk into an exam room with a breast cancer diagnosis, the immune-class architectural signature is already loud. The reading climbs monotonically with the pre-diagnostic interval — quiet near diagnosis, louder at five years, louder still at ten — and the longest-window effect size is among the strongest pre-diagnostic signals in the methylation literature. The result replicates independently in a second EPIC-Italy sub-cohort drawn on the same 450K platform with the same panel. A woman sitting in a clinic today, told she is fine, may already be carrying a readable cellular signature of a disease that will not declare itself for ten years. We have been telling people they are fine when their own bodies have been trying to say otherwise for a long time.
Sources: Severi 2014 (EPIC-Italy GSE51057, GSE51032) · Xu 2020 JNCI (Sister Study panel) · VAL-093 + VAL-096 (cookbook).
🌐
Multi-Cancer Systemic Drift
VAL-046 · Pancreatic, Lung, Colorectal, Prostate · 4 independent cohorts
Look at the same EPIC-Italy blood from a decade before diagnosis — but now from the per-tissue angle, asking which cells the drift is coming from. The answer is not what most people would guess. At ten-plus years before clinical breast cancer, the loudest tissue tiles are pancreatic, kidney, head-and-neck, and colon. The breast tile reads near-null at this window. The cellular system as a whole is showing wear earlier than it should for the customer's age, and a specific tissue tile lights up only in the months before clinical diagnosis. Both components replicate concordantly across two independent cohorts. The pre-diagnostic signal is not "early breast cancer" — it is a body that is already aging differently than its calendar age would predict, with the localized cancer arriving on top of that body-wide drift years later.
Sources: VAL-093 + VAL-096 (cookbook) · Loyfer 2023 array atlas (Stage 2 cell-of-origin reference) · Severi 2014 (EPIC-Italy).
📏
Organ-Wide Field Effect
VAL-037 + VAL-039 · 24 TCGA cancer types · n = 1,109 STN
"Adjacent normal" tissue in a pathology report is the surrounding tissue a surgeon takes as a clean control alongside the tumor itself. It is supposed to be healthy. The framework reads it differently. Across two dozen TCGA cancer types, adjacent-normal tissue sits noticeably above the reference of cells that have never been near a tumor — about a quarter of the way from healthy toward the tumor reading. Six studies that recorded distance from the tumor show a clean gradient: closer to the tumor, more drift; further away, less. Even five to ten centimeters out, the drift is still present. What clinicians have been calling a clean margin is, by this reading, an organ-wide architectural change that standard histopathology was never built to see. Whether that has implications for recurrence rates or for surgical margin guidelines is an open clinical question that pathologists, not the framework, will need to answer.
Sources: VAL-037 + VAL-039 (cookbook) · TCGA pan-cancer methylation reference · six distance-annotated tissue studies (lung, breast, colon, prostate, HCC, gastric).
Colorectal — Full Trajectory
Dark Blue ribbon · VAL-042 + VAL-044 + VAL-046
The colon adenoma-to-carcinoma sequence is the best-documented cancer progression in human medicine. The framework reads each clinical stage as a different reading on the same gauge — healthy colon, hyperplastic polyp, tubular adenoma, advanced adenoma with high-grade dysplasia, invasive cancer — and the gauge climbs monotonically through all five. Tubular adenoma crosses into the framework's pre-cancer tier; invasive disease crosses the floor. When a CRC patient goes through FOLFOX chemotherapy, responders show the gauge dropping back toward the healthy band and complete responders return to normal. Non-responders stay elevated. The framework also reads the tumor itself across two separate cell populations inside a single biopsy — the tumor cells against their own architectural floor, and the immune cells that have moved into the tumor against the immune-class panel. Colonoscopy raised colorectal cancer survival from fifty percent to sixty-five percent over four decades. It did not cure cancer. It moved the diagnosis earlier, when treatment still works. A blood-based cellular-state reading offers the same kind of opportunity, applied to people in the years before they become eligible for routine colonoscopy at all.
Sources: Hou 2012 Am J Epidemiol · Luo 2014 Gastroenterology · Parikh 2019 Nat Med · Reinert 2019 JAMA Oncol · VAL-061 + VAL-062 (TCGA-COAD HM450 paired tumor/adjacent-normal). ACS Colorectal Cancer Facts & Figures 2023-2025 (survival trend).
Breast — A Decade of Runway
Pink ribbon · Secretory class · VAL-047 + VAL-046 + VAL-044 + VAL-039
Mammography is the most successful cancer screening program in the history of American medicine. Five-year breast cancer survival rose from seventy-five percent in the mid-1970s to ninety-one percent today, and the change came not from curing cancer differently but from finding it earlier. The framework asks what additional runway is possible if the cellular signal can be read in blood years before any imaging modality could see the tumor.

The answer in two independent EPIC-Italy cohorts: a cellular signature is loudly present more than ten years before clinical diagnosis, climbs monotonically with the pre-diagnostic interval, and replicates concordantly across both cohorts. By the two years before diagnosis, the breast tissue tile itself starts to localize. About half of women who go on to develop breast cancer are already sitting in the upper decile of their age-matched healthy peers a full decade before their diagnosis, and the fraction climbs further as the diagnostic window approaches.

The same instrument addresses the dense-breast problem mammography cannot. Forty percent of American women have dense breast tissue, and mammographic sensitivity drops sharply in that population. The framework reads methylation entropy, not X-ray absorption — dense and non-dense tissue share the same architectural floor.

The screening did not cure cancer. It moved the diagnosis earlier, when treatment still works. The pre-diagnostic signal documented here is the cellular substrate of that same opportunity, with a decade of runway instead of two years, on a blood draw most patients are already getting for other reasons.
Sources: Severi 2014 (EPIC-Italy GSE51057, GSE51032) · Xu 2020 JNCI (Sister Study panel) · ACS Breast Cancer Facts & Figures 2024-2025 (survival trend) · VAL-093 + VAL-096 + VAL-060 (cookbook).
Pancreatic — The Largest Near-Term Runway
Purple ribbon · Secretory class · VAL-046 + VAL-007 + VAL-040
Pancreatic adenocarcinoma carries the lowest survival rate of any major cancer in the United States. Across all stages combined, five-year survival sits at thirteen percent. For patients diagnosed with metastatic disease — which is most patients at presentation, because the disease is typically silent until it has spread — five-year survival drops to three percent. The American Cancer Society projects approximately sixty-eight thousand new cases in the United States in 2026 with approximately fifty-three thousand deaths.

There is no recommended general-population screening test. The disease is usually caught when a patient develops jaundice, weight loss, or new-onset diabetes severe enough to trigger an imaging workup, by which point the window for curative surgery has typically closed. Stage IA pancreatic cancer caught in time for the Whipple procedure has five-year survival above thirty percent in real-world cohorts — a tenfold improvement over the all-stages baseline.

The framework's pancreatic application reads a cellular signature in pre-diagnostic blood drawn years before the cancer presents. The framework also flags a clinical pattern oncologists already recognize but currently have no instrumental screening for: new-onset Type 2 diabetes after age fifty paired with any directional cellular-departure signature is a paraneoplastic-PDAC workup trigger.

If even a fraction of pancreatic cancers were caught at the stage where the Whipple is curative instead of palliative, the change in absolute survival would be measured in tens of thousands of lives per year in the United States alone.
Sources: SEER + ACS 2026 mortality projections · VAL-046 (Rotterdam Study pre-diagnostic) · VAL-069 (TCGA-PAAD 324-CpG directional fallback panel) · VAL-007 (pancreatic exocrine cfDNA tissue-specific signal).
Glioma — LGG & GBM
Grey ribbon · Terminal class (tightest floor) · VAL-007 + VAL-044
Brain cancer has historically been undetectable in standard peripheral blood at array resolution. Cell-free DNA from neurons appears in the bloodstream at fractions so small that conventional methylation reference atlases were calibrated assuming the signal would not be reachable. The framework's earlier framing accepted this as a structural limitation — read brain disease in cerebrospinal fluid, not plasma.

That framing was wrong, and the reason is mundane: the older atlases used a bulk-tissue brain reference that did not separate cortical-neuron signal at the resolution standard methylation arrays measure. When a more recent atlas with a sorted-cell cortical-neuron entry is used as the Stage 2 reference, the signal in glioma plasma reads cleanly. Glioma plasma carries roughly four times more cortical-neuron cell-free DNA than healthy reference plasma, and the same effect appears in pre-surgery treatment-naive samples — meaning the signal is the disease, not the surgery or chemotherapy.

The same methodology run against Alzheimer's blood reads cortical-neuron cfDNA at near-null. AD does not elevate cortical-neuron cfDNA at this resolution, while glioma does. For a patient presenting with cognitive symptoms, the same instrument reads whether the cellular signature is consistent with neurodegeneration or with a tumor — from a single peripheral blood draw, with no need for a lumbar puncture, contrast MRI, or amyloid PET to start the differential.
Sources: VAL-090 (GSE180683 EPIC peripheral blood, n=76 glioma vs n=177 healthy) · VAL-091 (AD layered-atlas Stage 2 null) · Loyfer 2023 array atlas · Salas-Wiencke 2022.
Alzheimer's — Multi-Class at the Thermodynamic Level
Purple ribbon · Multi-class (4 elevated) · VAL-040
Alzheimer's disease has approximately 6.9 million prevalent cases in the United States today and is projected to roughly double by 2050 if no preventive intervention emerges. The new amyloid-targeting therapies — lecanemab and donanemab — are FDA-approved only for patients in mild cognitive impairment or mild dementia stages, with confirmed elevated brain amyloid. The drugs are not effective once the patient has progressed past mild dementia. The clinical bottleneck is identifying patients early enough that disease modification can do meaningful work, and current diagnostic pathways rely on cognitive testing that detects the disease only after symptoms have begun.

The framework reads AD as a systemic multi-class phenomenon at the cellular thermodynamic level — not a localized brain event. Four cell architectural classes show drift in AD cohorts: brain cortex itself, peripheral blood immune cells, pancreatic islets, and cerebral vasculature. Every tested tissue-class combination shows a severity gradient from early to late disease. The directional panel built for AD blood-based screening reads a real signal years before clinical presentation in independent cohorts.

A client identified as showing AD-pattern drift could enter monitoring earlier, begin pharmaceutical intervention earlier when those interventions still work, and have time to make the legal, financial, and family decisions that AD eventually forces. For some, that lead time may be the difference between writing the will yourself and having someone else write it for you. For some it may be the difference between being there for a grandchild's first communion and not being there.
Sources: De Jager 2014 Nat Neurosci (ROSMAP n=708) · Shireby 2022 Brain (BDR n=1,408) · Nabais 2021 Genome Biol (peripheral blood meta-analysis n=3,424) · Lunnon 2014 Nat Neurosci · VAL-040 + VAL-051 (cookbook).
🧭
Tissue-of-Origin Localization
VAL-041 · 10 cancer types · 10/10 top-1 correct
A standard plasma sample contains cell-free DNA fragments from every tissue in the body simultaneously, mixed together. Reading the bulk plasma against any single tissue's architectural floor is a category error — the math does not work because the signal is averaged across cells the architecture is not for. The fix is to first separate the bulk into per-tissue fractions using a published methylation atlas, then score each separated fraction against the architectural floor of its own class. When this is done correctly, the framework localizes the tissue of origin correctly in every cancer type tested. The current implementation runs a layered atlas — the original Moss 2018 reference, plus the more recent Loyfer 2023 array atlas that supplies the cortical-neuron resolution that opened brain cancer in plasma. Stage 2 of the pipeline is what turns a routine annual blood draw into a tissue-resolved cellular-state reading, on the same sample most patients are already giving for cholesterol and metabolic panels.
Sources: Moss 2018 Nat Commun · Loyfer 2023 Nature (array atlas) · Liu 2020 Ann Oncol (extended marker panel) · VAL-041 (cookbook).
Seminoma — The Inversion Signature
Orchid ribbon · Pluripotent class · VAL-045
Most cancers read on the framework's gauge as elevation above the architectural floor — a cell is doing more methylation work than it should be. Seminoma and embryonal carcinoma read in the opposite direction. The pluripotent cells these tumors arise from sit, in their healthy state, near the upper limit of methylation entropy that any somatic cell can reach. There is nowhere up to go. When these cells transform, the methylation pattern moves down toward hypomethylation, away from the floor in the opposite direction from every other cancer the framework has been calibrated against. A detector built only to flag elevation would miss the entire signal — and roughly five thousand testicular germ cell tumor cases occur annually in the United States, the seminoma fraction making up about sixty percent of them. Reading in both directions by construction is the difference between catching those five thousand patients and missing them.
Sources: Shen 2018 Cell · Killian 2016 Cell Rep (seminoma hypomethylation) · VAL-045 (cookbook).
🐕
Cross-Species Validation — Canine Oncology
VAL-043 · 5 canine cancers · n = 104 Labradors
If the framework is reading a real physics quantity rather than a human-cohort artifact, the same architectural floors should appear in non-human mammals. The test ran five canine cancers against the same class floors calibrated on human tissue, and the cross-species offset between dog and human readings is essentially zero across seventy million years of evolutionary divergence. Dog aging tracks human aging on the same gauge with near-perfect correlation. Dogs age roughly seven times faster than people, which means a prospective canine cohort can deliver a longitudinal validation in eighteen months that a human cohort would take a decade to produce. Veterinary oncologists and the dogs in their clinics are how the next layer of evidence becomes available before the human cohorts catch up — and in the process, the same instrument helps the dogs.
Sources: Wang 2020 Cell Reports (Labradors) · Pal 2016 Cancer Res · Beck 2020 Vet Comp Oncol · Decker 2015 PLoS Genet · VAL-043 (cookbook).
📉
Treatment Response Trajectory
VAL-044 · 5/5 clinical trials · GBM, CRC, BRCA, AML, melanoma
The same gauge that climbs as a cell population drifts away from healthy can drop again when treatment works. Across five tested clinical trial cohorts — glioblastoma on the Stupp protocol, colorectal cancer on FOLFOX, breast cancer in adjuvant chemotherapy, acute myeloid leukemia on 7+3 induction, and melanoma on anti-PD-1 — the framework's reading separates patients who respond to therapy from patients who do not. Responders show the gauge dropping back toward the healthy band. Complete responders return to normal. Non-responders stay elevated or continue to rise, and the divergence is visible in serial samples drawn during treatment.

Acute myeloid leukemia shows the strongest single-cohort signal in the framework's catalog. AML cancer cells are the same myeloid-lineage cells the universal Stage 1 panel was built to read, which means the framework reads the disease cells directly rather than against an immune-cell background. The signal at diagnosis is essentially impossible to miss.

Architectural drift precedes the disease. Architectural recovery accompanies the treatment that works. Both are measurable in the same blood draw — which means a patient on therapy, and the oncologist managing that therapy, can see whether the treatment is working months before standard imaging would show it.
Sources: Ceccarelli 2016 Cell + Fan 2021 J Neurooncol (GBM) · Parikh 2019 Nat Med (CRC) · Stover 2018 J Clin Oncol (BRCA) · Ley 2010 NEJM + Ding 2012 Nature (AML) · Cabel 2018 Ann Oncol (melanoma) · Glass 2017 (GSE62298 AML) · VAL-044 + VAL-082 (cookbook).
Prostate — PIN to Metastatic
Light Blue ribbon · Secretory class · VAL-042 + VAL-046
Prostate cancer screening has been arguing with itself for thirty years. PSA testing flags many men whose biopsies turn up nothing dangerous, and the concern about overtreatment has shifted clinical guidelines back and forth more than once. The clinical question is not whether a screening test can detect — it is whether a screening test can stratify, telling clinicians which men in the gray zone actually need workup and which can safely watch and wait.

The framework reads the full clinical progression — healthy prostate, low-grade prostatic intraepithelial neoplasia, high-grade PIN, Gleason 3+3, Gleason 4+3 and higher, metastatic castrate-resistant — as six distinct readings on the same gauge, each higher than the last. The gauge crosses into the pre-cancer band at HGPIN and breaches the floor at clinically significant disease. In a separate prospective cohort, blood drawn three years before prostate cancer diagnosis already shows immune-class drift. Prostate-epic was also the first card in the framework's catalog built tissue-first, on a published cohort of African-American men where direct tissue measurement carries clear signal.

Roughly one in five men who undergo radical prostatectomy develop long-term urinary incontinence, and roughly two in three experience long-term erectile dysfunction. The clinical cost of treating the wrong prostate cancer is high. A reading that helps the urologist decide which man in the PSA gray zone actually needs the biopsy — and which one can watch and wait another year — is the kind of information the current toolkit does not have.
Sources: Jerónimo 2008 Clin Cancer Res · Aryee 2013 Sci Transl Med · Damaschke 2017 (zonal gradient) · Horvath 2014 Health ABC · Berglund/Yamoah/Kresovich 2024 (GSE269244, n=238 African-American men, EPIC 850K) · VAL-042 + VAL-046 + VAL-058 (cookbook).
Complete Validation Record — VAL-001 through VAL-128
128+
sealed validation studies executed
24+
cancer types tested across the cascade
4,304
matched tumor-normal pairs (TCGA pan-cancer)
30
TCGA cancer types · 29/30 confirmed
10+ yr
breast pre-diagnostic window (VAL-047)
0
free parameters · zero cancer training data
The full testing record at a glance. The methylation core (VAL-001 to VAL-013) established the pan-cancer A-score on TCGA, Health ABC, Roadmap Epigenomics, Moss 2018 cfDNA atlas, OSK reprogramming, DunedinPACE, pan-mammalian aging, and Labrador cross-species. The multimodal expansion (VAL-014 to VAL-033) added the five-substrate framework across nucleosome occupancy, fuzziness, WPS, and fragment-size alongside methylation. The vertebrate extension (VAL-034 to VAL-036) confirmed pan-mammalian lifespan correlation and ectotherm temperature correction. The multi-class drift cascade (VAL-037 to VAL-046) added organ-wide field effect, tissue-of-origin localization, pre-cancer progression, treatment response, inversion specificity, cross-species replication, multi-class AD drift, and the pre-diagnostic multi-cohort signature. VAL-047 closed the cascade loop with external per-patient 450K methylation on 1,581 real samples from three GEO cohorts.

The per-card tissue and pattern wave (VAL-048 to VAL-100) added card-specific tumor-vs-adjacent-normal validation on public TCGA cohorts — moving breast, lung, hcc, prostate, and crc to tissue_arm_validated tier and surfacing CRC's blood-vs-tumor-TIL compartment-direction-flip. The pancreatic-epic 324-CpG directional fallback panel established the pooled-null + directional-pass pattern as the second case alongside AD (TCGA-PAAD d = +1.51, 7/7 patients positive). Cervical-epic v0.1 documented the first opposite-sign Cohen's d across independent cohorts of the same disease, creating the exploratory_with_cohort_heterogeneity tier. Layered-atlas Stage 2 demonstrated glioma plasma reads cortical-neuron cfDNA at d = +1.96 — overturning the earlier prediction that plasma fails for terminal class — and confirmed AD does not elevate cortical-neuron cfDNA at array-NNLS resolution, enabling glioma-vs-AD differential diagnosis from a single IDAT. The first multi-cohort run-everything pass surfaced replicating sub-cell-type resolution gains in breast pre-diagnostic blood (UniLIFE 19-cell aTreg at >10yr, aBnv at 0-2yr).

The multi-atlas calibration discipline wave (VAL-101 to VAL-128) anchored every active card on a structurally-separated TCGA n=210 healthy calibration cohort with per-atlas q5 thresholds sealed before any disease scoring. Cardio-epic v0.3 (VAL-108-113), prostate-epic v0.3 with the ProstateRef luminal dedifferentiation pattern at d_paired = −0.767 (VAL-117-118), bladder-epic v0.1 with the cohort-substrate-coverage gate (VAL-119-122), and gastric-esophageal-epic v0.1 with within-cancer histological-subtype discrimination at d_ESCC−EAC = −1.06 (VAL-123-128). The wave produced 10 new cross-card lessons and 13+ new pre-flight checklist gates. All 128+ sealed VAL scripts and their JSON pass/fail records are in the public repository. Pre-clinical research only; prospective clinical validation required before individual-patient use.
EDEAR · The discipline behind the development

Every prediction sealed.
Every threshold timestamped.

The framework underlying EDEAR is being built specifically to make data-fitting impossible. Every prediction has a timestamp on the public IAM-Validation GitHub repository. Every threshold has a sealed SHA-256 hash. Every reference atlas in the production vault has a public provenance trail back to the published paper that produced it. The discipline that surrounds the science is not optional plumbing — it is what allows the claims to be claims at all. EDEAR is pre-clinical research. Prospective clinical validation has not yet been performed. The work in this section is the foundation that has to be in place before a clinical pilot can honestly be run.

Designed · Currently being mapped
IAMAtlas — the first CpG atlas derived from first principles
While existing methylation panels are selected by training to discriminate one disease from healthy on one cohort, IAMAtlas was designed by deriving per-CpG informativeness posteriors directly from the architectural-class floor structure that anchors the entire framework — the same MCMC machinery cosmology uses to map the Cosmic Microwave Background, and the same machinery that calibrated the eight cell-class architecture values, applied to per-CpG informativeness across the methylation array. Marker selection becomes a derivation, not an import. The atlas is currently being mapped as we speak.
📊 What the client sees on their report
Six tiers, both directions, calibrated to physics.
Every cell architecture class on every report is scored against thermodynamic floors derived from physics. The client’s position is reported on a six-tier scale that runs in both directions — below normal for chronic immunosuppression, post-chemotherapy states, advanced disease with marrow infiltration; above normal for tissues running hot under chronic-driver pressure. The scale is the same for every disease card in the cookbook.
Below Normal
Normal
Marginal
Detectable
Urgent
Floor Breach
Class running cold — immunosuppression, post-chemo, marrow infiltration, cachexia.
At healthy operating temperature for the cell class.
Edge of healthy range — flag for trajectory.
Class running hot — clinical attention warranted.
Significant departure from healthy — act now.
Architectural floor crossed — cell class can no longer maintain identity.
The gauge runs in both directions because chronic-driver pressure pushes some tissue classes hot (cancer pre-diagnostic, autoimmune amplification, IBD active disease) and other clinical states push them cold (chronic immunosuppression, post-chemotherapy, advanced disease with marrow infiltration). EDEAR reports both because both are real clinical states. A client at NORMAL on every cell class is at the healthy operating temperature for her body. A client at MARGINAL on the immune class with a documented chronic driver in her history is the case the framework was built for.
🧬 The eight cellular architecture classes
One law of physics. Eight thermodynamic floors. Every human cell sits on one of them.
Every cell architecture class has its own Hmin(class) — the minimum methylation entropy consistent with that cell’s identity, derived from the same physics that governs gravitational decoherence and the Hubble horizon. The class-specific floors below were calibrated against published healthy-tissue reference data and frozen on April 6, 2026 with a SHA-256 hash on the public IAM-Validation repository. Every cancer card in the cookbook is built on the floor for the cell class the cancer originates from. Every disease card reads against the floor for the architecture class the disease lives in.
Architecture class
Hmin(class)
Example tissues
Disease cards reading this class
Terminal
Lowest ceiling in biology
0.77####
Neurons, cardiomyocytes
Alzheimer’s · Parkinson’s · LGG / GBM · cardiovascular trajectories
Immune
Read on every blood draw
0.83####
T cells · B cells · NK cells · neutrophils · monocytes
All cancer pre-diagnostic signatures · AML · CRC · bladder · gastric/esophageal · Crohn’s amendment
Secretory
Glandular & secretory epithelium
0.84####
Breast ducts · pancreatic islets · hepatocytes · prostate luminal
Breast · pancreatic · liver / HCC · prostate
Progenitor
Transit-amplifying cells
0.85####
Crypt-base columnar cells · intermediate-commitment cells · basal layer
Cervical CIN stratification · pre-cancer monitoring
Cycling
Renewing epithelium
0.85####
Colorectal · gastric · lung bronchial · bladder urothelium · skin
Colorectal · lung · bladder · gastric · esophageal
Stromal
Connective & vascular
0.86####
Fibroblasts · endothelial cells · mesenchymal · vascular smooth muscle
Cardiovascular trajectories · tumor microenvironment readout
Stem (adult)
Tissue-resident stem cells
0.87####
Hematopoietic stem · neural stem · intestinal stem · epidermal stem
Aging trajectories · stem-cell exhaustion endpoints · CHIP → MDS → AML cascade
Stem (pluripotent)
Highest ceiling — maximally open chromatin
0.98####
Embryonic stem cells · iPSC · germ cells
Reference architecture (developmental research)
The Hmin(class) values shown are redacted to two decimal places — the last four decimals are proprietary to Astro-GAPE calibration and were frozen with a SHA-256 hash on April 6, 2026, before any cancer cohort was scored against them. The full-precision values are protected. The redacted form is sufficient to verify that eight class-specific floors exist, that they span the predicted range from the lowest ceiling in biology (terminal) to the highest (pluripotent stem), and that they were locked before the validation evidence was generated.
🔬 What the cookbook covers
Seventeen clinical cards have been built and sealed against published, externally reproducible cohorts: breast (TCGA-BRCA + EPIC-Italy pre-diagnostic), Alzheimer’s (AIBL + AddNeuroMed + Religious Orders Study), colorectal (TCGA-COAD + TCGA-READ paired), lung (TCGA-LUAD smoking-stratified), liver and HCC (TCGA-LIHC + plasma cfDNA dose-response), AML (GSE62298), glioma (GSE180683 EPIC peripheral blood), pancreatic (TCGA-PAAD + Rotterdam Study pre-diagnostic), prostate (TCGA-PRAD + GSE269244 African-American), cervical (Verlaat Amsterdam CIN-stratified), cardiovascular (PAH + BAV + stroke etiology), bladder (TCGA-BLCA paired), gastric and esophageal with within-cohort histological-subtype discrimination (TCGA-STAD + TCGA-ESCA), and a Crohn’s disease pathway amendment for IBD blood-substrate readout (GSE87650). Every cohort, every cohort size, every Cohen’s d, every sealed prereg SHA-256 is on the public IAM-Validation GitHub repository and can be re-run by any skeptical reader from the source.
🧪 Why the discipline matters
The pre-registration discipline matters most when a prediction fails. Several have, openly. A pre-locked prediction that the universal Stage 1 immune panel would elevate on chronic inflammatory disease returned a clean null — and the failure clarified that Stage 1 is a tissue and cancer signal, not a generic chronic-inflammation marker. A pre-locked prediction that sorting blood cells from a Crohn’s disease cohort would amplify the within-cell-type signal returned the opposite direction — and the failure re-framed what every immune deconvolution atlas in the cookbook actually measures. None of these failures were quietly walked back. Each one was sealed, dated, and published as the lesson it was. Each one made the framework smaller in scope, more specific in claim, and more honest about its own boundaries.
Run everything, every time
One blood draw, one Illumina EPIC array, one workflow. Every IDAT runs through every panel and every reference atlas in the production vault, in parallel. The client’s report displays anomalies and collapses uninformative tiles for readability; the underlying scoring is exhaustive. No conditional gating. No disease-specific Stage 1 substrate. The instrument is designed to look at every cell class and every tissue, every time, and the disease cards interpret what shows up.
Trajectory, not coordinate
A single thermometer reading is a coordinate. A pair of readings six months apart is a rate of change. The intended deployment model is serial sampling — annual or biannual blood draws layered onto routine care — converting what would otherwise be a single coordinate into a personal trajectory unique to the customer, scored against her own multi-year baseline rather than a population average. That is the difference between reactive and proactive medicine.
📄 Read the Operations Summary 🔓 Open GitHub Repository Cookbook validation source open to any reader · Atlas vault SHA-anchored · 128+ sealed VAL studies

What is the Informational
Actualization Model?

IAM describes the engine underneath all phenomena — the mechanism by which potential becomes actual. Every irreversible physical transition pays a thermodynamic cost at the nearest encoding surface. IAM extends that accounting across cosmic time to explain the expansion of the universe — and measured at the scale of a cell nucleus, defines what healthy means from first principles. One engine. Every scale. The same bill.

From the formation of cosmic structure to the behavior of a qubit, from the switching of a transistor to the failure of cellular control, IAM treats each irreversible transition as a physical event that must be paid for at the nearest encoding surface. Different substrates expose that cost differently, but the common structure is the same.

That is why the same framework can be used to study cosmology, quantum hardware, semiconductor scaling, and cellular thermodynamics without pretending the domains are identical. The substrate changes. The noise mechanism changes. The measurable inputs change. The underlying thermodynamic architecture does not.

IAM is the derivational base of the broader research program and the foundation of the three IAMPerformance instruments — Astro-GAPE, Astro-QAPE, and Astro-SCAPE. Astro-GAPE is open science. Astro-QAPE and Astro-SCAPE are commercial instruments.

IAM's Law
Every irreversible transition pays a thermodynamic cost to the nearest encoding surface.
The cost is set by the substrate’s operating temperature and the information produced. It is irreversible and permanent — it cannot be recovered, reversed, or engineered away. Different substrates expose the cost differently, but the underlying law is the same. The quantitative form and the per-substrate calibration are proprietary to the IAMPerformance Architecture-class framework (patent pending US 64/012,720 and US 64/014,568).
Malignancy: a cell that has exceeded its thermodynamic maintenance budget.
Black hole: a star that did the same.
A star accumulates entropy through nuclear burning. When its gravitational encoding surface reaches capacity, the star can no longer pay the cost of remaining itself. The black hole is not a thing that formed — it is what remains when a gravitational system can no longer maintain its thermodynamic identity. The event horizon is the encoding surface that absorbed what the star could not sustain.

A cell accumulates entropy through failed methylation maintenance. When the A-score crosses the architecture floor, the cell can no longer pay the cost of remaining itself. The malignant phenotype — invasion, proliferation, metabolic shift — is not the failure. It is the consequence of the failure. The physics gives you the signal before the consequence arrives.

Both are floor breaches. Both have encoding surfaces. Both were predictable before the event. The black hole is the FLOOR BREACH tier of stellar architecture. Same equation. 37 orders of magnitude apart.
The dimensionless ratio that measures how far any system sits above saturation is the Mahaffey Number. See the Common Thread section below.
Patent pending · 64/012,720 & 64/014,568

One law, one fingerprint.

Every domain has a physics floor, an architecture ceiling, and a dimensionless ratio measuring the distance between them. The Mahaffey Number is that ratio — IAMPerformance’s dimensionless metric for how far any system sits above its physics floor and below its architecture ceiling. When the Mahaffey Number reaches the architecture-specific ceiling, the system transitions. The table below covers every architecture class from the gravitational horizon to the human cell nucleus. The quantitative form and per-domain calibration are proprietary to IAMPerformance.

Cosmology
Structure formation as thermodynamic load
The original IAM program treats irreversible information production from gravitational structure formation as a real thermodynamic contribution to the entropy budget. The result: a horizon-based accounting in which the cosmological constant, baryon asymmetry, and expansion history follow from one internally consistent framework — zero free parameters beyond Planck 2018.
Astro-QAPE
The qubit operating range
Every quantum architecture has a material floor (irreducible gate error from the substrate) and an architecture ceiling (substrate inversion — where T1 hits the material limit). Astro-QAPE places every published platform within that range, derives both boundaries from first principles, and quantifies how much room remains. The QEC threshold is a milestone inside the range — not the ceiling.
Astro-SCAPE
The chip operating range
Every ISA class has a physics floor (irreducible switching cost at operating temperature) and an architecture ceiling (the ISA overhead floor — proprietary per-ISA multiples in Astro-SCAPE calibration). Astro-SCAPE places every published chip within that range. The Dennard wall is the ceiling: node shrinks keep delivering until the ISA overhead floor is reached. It cannot be crossed without ISA redesign.
Astro-GAPE
The cell operating range
Every cell architecture class has a physics floor (irreducible methylation maintenance cost at 37°C) and a biology ceiling (Hmin(class) — the minimum methylation entropy consistent with that cell's identity). Astro-GAPE places a cell's epigenomic state within that range from a single methylation measurement. Above the ceiling: malignant transformation. Pre-clinical research only. Not intended to diagnose any individual.
The Mahaffey Number — IAMPerformance’s dimensionless operating-range metric
One ratio. From the cosmic horizon to the human cell nucleus — 37 orders of magnitude, one law.
From the cosmic horizon to the transistor junction to the human cell nucleus — one framework, three commercial instruments.

Each domain has its own operating range. The Mahaffey Number normalizes any platform onto the same 0-to-1 scale inside its own range, allowing cross-domain comparisons of where every system sits between the irreducible physical minimum and the architecture-specific ceiling. The quantitative form and per-domain calibration are proprietary to IAMPerformance (patent pending US 64/012,720 and US 64/014,568).

The floor, the ceiling, and what crossing the ceiling means
The floor — ℳnorm = 0
The universal physical minimum for an irreversible information event at the local temperature. The same floor for every transistor ever built, every qubit ever measured, every cell ever divided. No engineering crosses it downward.
The ceiling — ℳnorm = 1.0
The architecture-specific physics limit — where normal operation ends. The ceiling is different for every architecture class. For qubits: substrate inversion. For chips: ISA overhead floor. For cells: Hmin(class). For black holes: the Schwarzschild condition. All derived, not fitted.
Crossing the ceiling — ℳnorm > 1.0
The system does not degrade gradually. It transitions. In semiconductors: the Dennard wall — frequency stagnates, power density explodes. In qubits: the Substrate Inversion — material ceiling reached, engineering returns zero. In cells: malignant transformation — epigenomic identity lost. In gravity: a black hole forms. Note: performance milestones like the QEC threshold in quantum computing are targets inside the operating range — not ceilings.
What the floor and ceiling mean in each domain
Semiconductor (Astro-SCAPE)
Floor: The irreducible physical minimum for a transistor switch at operating temperature. Identical for every chip at every foundry. Cannot be engineered away.
Ceiling: ISA overhead floor — architecture-specific, proprietary to Astro-SCAPE calibration. Set by instruction set architecture, not by material. Cannot be lowered without ISA redesign.
Crossing: Dennard breakdown — single-core frequency stagnation, power density explosion, end of voltage scaling.
Quantum computing (Astro-QAPE)
Floor: Material A-floor — the irreducible gate error for that substrate. Proprietary to Astro-QAPE calibration.
Ceiling: Architecture physics limit — substrate inversion for SC qubits, motional heating saturation for ion traps, BBR depopulation for Rydberg. Normal operation ends here.
Crossing: Coherence budget exhausted — engineering returns zero. Requires new substrate or new gate mechanism.
The QEC threshold (p < 10⁻³) is a performance milestone inside the operating range — not a physics ceiling.
Cellular (Astro-GAPE)
Floor: The irreducible physical minimum for a methylation maintenance event at body temperature. Universal — the same floor for every cell in every human body. Cannot be reduced by any intervention.
Ceiling: Hmin(class) — the minimum methylation entropy for that cell architecture class. Different for each of the eight somatic cell classes, proprietary to Astro-GAPE calibration. When a client enters their methylation score into Astro-GAPE, the output is their ℳnorm — how close they are to their class ceiling.
Crossing: Malignant transformation — epigenomic identity lost, architecture class can no longer be maintained.
Gravitational
Floor: Zero-point vacuum — set by the geometry of spacetime itself.
Ceiling: Horizon saturation — the gravitational encoding surface reaches capacity. Derived from first principles — not fitted.
Crossing: Black hole formation — a new encoding surface appears.
ℳ operating range — all architecture classes and platforms, telescope to microscope
Architecture / platform
n
Floor  ·  Ceiling  ·  What ceiling means
norm
Room remaining
norm = (current − floor) / (ceiling − floor).  0 = at the physics floor (irreducible minimum — best achievable).  1.0 = at the physics ceiling (normal operation ends — architecture transition).  Values above 1.0 mean the ceiling has been crossed.  The QEC threshold in quantum computing is a performance milestone inside the operating range, not a physics boundary.
Gravitational & Cosmological — Floor = zero-point vacuum  |  Ceiling = horizon saturation
Black hole horizon
Any collapsing mass
1/(4ℓ²)
Floor: Hawking zero-point vacuum  Ceiling: Schwarzschild — SBH fills horizon capacity  → New encoding surface forms
1.00
at ceiling — transition
IAM cosmological
Structure formation — today
5/2
Floor: Electroweak transition a=0  Ceiling: ℰ(a)=e — full matter actualization, never reached  → Expansion suppression
0.37
63% room (asymptotic)
Quantum computing (Astro-QAPE) — Floor = material A-floor  |  Ceiling = coherence / control saturation for that architecture class  |  QEC threshold is a milestone inside the range
Oxford Ionics EQC
Ion B Ca+  |  p=8.4×10⁻⁵  |  ℳnorm=0.056
[proprietary]
Floor: 3×10⁻⁵ Ca+ electronic state noise  Ceiling: ~1×10⁻³ RF control bandwidth saturation  → Control correction fails
● QEC milestone: already crossed (p < 1×10⁻³) — fault-tolerant operation achievable
0.056
94% room
Quantinuum Helios 98Q
Ion A Ba+  |  p=7.9×10⁻⁴  |  ℳnorm=0.141
[proprietary]
Floor: 1×10⁻⁴ Ba+ motional heating limit  Ceiling: ~5×10⁻³ motional heating saturation  → Sympathetic cooling fails
● QEC milestone: crossed (p < 1×10⁻³)
0.141
86% room
Quantinuum H1-1
Ion A Ba+  |  p=8.6×10⁻⁴  |  ℳnorm=0.155
[proprietary]
Floor: 1×10⁻⁴  Ceiling: ~5×10⁻³ motional saturation  → Sympathetic cooling fails
● QEC milestone: crossed
0.155
85% room
Quantinuum H2-1 56Q
Ion A Ba+  |  p=1.84×10⁻³  |  ℳnorm=0.355
[proprietary]
Floor: 1×10⁻⁴  Ceiling: ~5×10⁻³ motional saturation  → Operating normally — 36% of range used
◯ QEC milestone: not yet reached (p > 1×10⁻³) — NISQ regime
0.355
65% room
Google Willow 105Q
SC CZ Al+PR  |  p=1.5×10⁻³  |  ℳnorm=0.125
[proprietary]
Floor: 1×10⁻³ Al+PR TLS density  Ceiling: ~5×10⁻³ TLS saturation regime  → Decoherence uncontrolled
◯ QEC milestone: not yet reached — NISQ regime
0.125
88% room
IBM Heron R2 156Q
SC CZ Nb+AlOx  |  p=2.0×10⁻³  |  ℳnorm=0.835
[proprietary]
Floor: 1×10⁻³ Nb+AlOx TLS  Ceiling: ~2.2×10⁻³ substrate inversion (T1=300μs, T1*=325μs)  → Material ceiling approaching
◯ QEC milestone: not yet reached — NISQ regime
0.835
17% room
IBM Nighthawk 120Q
SC CZ Nb+AlOx  |  p=2.154×10⁻³  |  ℳnorm=0.964
[proprietary]
Floor: 1×10⁻³  Ceiling: ~2.2×10⁻³ substrate inversion (T1=350μs > T1*=325μs)  → At material ceiling
△ Substrate inversion: T1 exceeds T1* — material ceiling reached
0.964
4% room — ceiling
IonQ Forte 36Q
Ion A Yb+  |  p=4.0×10⁻³  |  ℳnorm=0.789
[proprietary]
Floor: 3×10⁻⁴ Yb+ motional limit  Ceiling: ~5×10⁻³ motional saturation  → Operating normally — 79% of range used
◯ QEC milestone: not yet reached — class stall on Yb+ (Ba+ path needed)
0.789
21% room
Rigetti Cepheus-1 36Q
SC CZ Al/AlOx  |  p=4.0×10⁻³  |  ℳnorm=0.144
[proprietary]
Floor: 3×10⁻³ Al/AlOx TLS  Ceiling: ~1×10⁻² TLS saturation  → Operating normally
◯ QEC milestone: not yet reached
0.144
86% room
Rigetti Cepheus-1 108Q
SC CZ Al/AlOx chiplet  |  p=1.0×10⁻²
[proprietary]
Floor: 3×10⁻³ Al/AlOx  Ceiling: ~1×10⁻²  → At TLS saturation ceiling
△ At ceiling — chiplet interconnect overhead exhausting coherence budget
1.005
at ceiling
Rigetti Ankaa-3 84Q
SC ECR Al/AlOx  |  p=1.0×10⁻²
[proprietary]
Floor: 3×10⁻³ Al/AlOx  Ceiling: ~1×10⁻²  → At TLS saturation ceiling
△ At ceiling — Al/AlOx TLS density limiting ECR coherence
1.005
at ceiling
QuEra Gemini 260Q
Neutral Rydberg Rb  |  p=8.0×10⁻³  |  ℳnorm=0.202
0.351
Floor: 5×10⁻³ Rydberg state lifetime noise  Ceiling: ~2×10⁻² BBR-driven Rydberg depopulation  → Blockade coherence lost
◯ QEC milestone (5×10⁻³): not yet reached for this class
0.202
80% room
Microsoft Majorana 1
Topological SC  |  gate data pending
[proprietary]
Floor and ceiling pending first gate publication. Class assigned provisionally as topological SC.
TBD
awaiting data
Google Neutral Atom 2026+
Neutral Rydberg Rb  |  gate data pending
0.351
Same floor and ceiling as QuEra Gemini class. First gate publication will place it in the range.
TBD
awaiting data
Semiconductor (Astro-SCAPE) — Floor = physical minimum per switch  |  Ceiling = ISA overhead floor (architecture physics limit)  |  Engineering brings Astro-SCAPE value down toward ceiling
AMD Ryzen 9 9950X
Zen5 4nm  |  Astro-SCAPE 576×  |  8.2× above ceiling
[proprietary]
Floor: physical minimum per switch (universal)  Ceiling: x86 ISA floor 70× minimum  → Cannot improve below without ISA redesign
8.2×
12% from ceiling
Intel Core Ultra 9 285K
Arrow Lake 3nm  |  Astro-SCAPE 570×  |  8.1× above ceiling
[proprietary]
Floor: physical minimum  Ceiling: x86 ISA floor 70×  → Same ceiling as AMD — different temperature sensitivity
8.1×
12% from ceiling
Apple M5
TSMC N3P  |  Astro-SCAPE 66×  |  4.4× above ceiling
0.794
Floor: physical minimum  Ceiling: Apple ARM ISA floor 15× (most efficient ISA class)  → Dennard wall
4.4×
23% from ceiling
Apple M4
TSMC N3E  |  Astro-SCAPE 68×  |  4.5× above ceiling
0.794
Floor: physical minimum  Ceiling: Apple ARM ISA floor 15×  → Dennard wall
4.5×
22% from ceiling
Apple M3
TSMC N3B  |  Astro-SCAPE 65×  |  4.3× above ceiling
0.794
Floor: physical minimum  Ceiling: Apple ARM ISA floor 15×  → Dennard wall
4.3×
23% from ceiling
Qualcomm Snapdragon X Elite
TSMC 4nm  |  Astro-SCAPE 316×  |  9.0× above ceiling
0.500
Floor: physical minimum  Ceiling: Qualcomm ARM mobile ISA floor 35×  → Dennard wall
9.0×
11% from ceiling
NVIDIA B200 SXM
TSMC 4NP dual-die  |  Astro-SCAPE 641×  |  2.6× above ceiling
0.650
Floor: physical minimum  Ceiling: GPU ISA floor 250× (parallel compute overhead)  → Dennard wall
2.6×
39% from ceiling
NVIDIA H200 SXM
Hopper 4nm  |  Astro-SCAPE 1,326×  |  5.3× above ceiling
0.650
Floor: physical minimum  Ceiling: GPU ISA floor 250×  → Dennard wall
5.3×
19% from ceiling
NVIDIA H100 SXM5
Hopper 4nm  |  Astro-SCAPE 1,435×  |  5.7× above ceiling
0.650
Floor: physical minimum  Ceiling: GPU ISA floor 250×  → Dennard wall
5.7×
17% from ceiling
Cellular (Astro-GAPE) — Floor = physical minimum per methylation maintenance event  |  Ceiling = Hmin(class) — minimum entropy for that cell architecture  |  Crossing ceiling = malignant transformation
Cellular — 8 architecture classes
Terminal · Immune · Secretory · Progenitor · Cycling · Stromal · Stem (adult) · Stem (pluripotent)
[proprietary]
Floor: Universal physical minimum per CpG  Ceiling: Hmin(class) — class-specific minimum entropy [proprietary]  → Ceiling crossing = malignant transformation, neurodegeneration, end-stage disease
➤ Patient-facing tier vocabulary & full architecture-class table: EDEAR section below
see EDEAR
8 classes · 17 cards built
Quantum computing ℳnorm = (Acurrent − Afloor) / (Aceiling − Afloor) where Aceiling is the architecture physics limit (motional saturation / substrate inversion / BBR depopulation). ● = QEC milestone crossed (fault-tolerant operation achievable). ◯ = QEC milestone not yet reached (NISQ regime). The QEC threshold is a performance milestone inside the operating range, not a physics boundary.   Semiconductor ℳnorm shown as current/ceiling ratio (how many times above the ISA floor ceiling). Engineering reduces this ratio toward 1.0× (at ceiling).   Cellular ℳnorm = (H(β) − Hminglobal) / (Hmin(class) − Hminglobal). All cellular positions from published Roadmap Epigenomics reference data for healthy cells.
A technical reader does not have to accept every derivation at once to understand the unifying thesis: the same thermodynamic architecture is leaving the same fingerprint in every system that processes irreversible information. The Mahaffey Number is that fingerprint quantified. The three APE instruments are Mahaffey Number meters — each one measuring the same ratio in a different domain, in the domain-appropriate units, from first principles.

The APE Family
Astro-GAPE · Astro-QAPE · Astro-SCAPE — one law, three instruments, one fingerprint

How we operate

IAMPerformance runs the Architecture-class framework on its own infrastructure. Clients submit platform parameters and receive structured reports showing the Mahaffey Number position inside the physics-derived operating range, plus trajectory and competitive analysis. The framework’s quantitative form and per-domain calibration are proprietary and remain with IAMPerformance. Customers see outputs, not internals. Patent pending US 64/012,720 and US 64/014,568.

APE
Architecture-class Performance Engine Family

Each APE is a domain-specific Mahaffey Number meter — measuring how far each architecture sits above its physics floor and below its architecture ceiling. The same dimensionless ratio. The same first-principles derivation. Different substrates, different ceilings, different inputs — but the same underlying law. The Architecture-class framework is proprietary to IAMPerformance.

Published Engines
● Published · Issue 002
Astro-QAPE
Astro-Quantum Architecture-class Performance Engine
Quantum computing hardware — gate fidelity, material floor, architecture class, substrate inversion ceiling, competitive trajectory. Input: one published gate error rate.
● Published · Issue 002
Astro-SCAPE
Astro-Semiconductor Architecture-class Performance Engine
Semiconductor & integrated circuit performance — ISA floor, Dennard proximity, three-component decomposition, node scaling regime, competitive trajectory. Input: TDP, transistors, frequency.
● Open Science · Issue 002
Astro-GAPE
Astro-Genetic Architecture-class Performance Engine
Cellular architecture thermodynamics — epigenomic floor, Hmin ceiling, cancer research, aging trajectory. Input: your methylation β from a tissue-class-specific blood panel. Zero training data. Pre-clinical research only — not intended to diagnose any individual.
○ Framework scope
APE+
Additional domains
Neural implants · DNA storage · Solid electrolytes · Neuromorphic. Same physics. Not yet built.
How it works — one number, derived from physics
01
Every domain has a Mahaffey Number For a qubit: how far above the material floor the gate error sits. For a chip: how far above the ISA overhead floor the switching efficiency sits. For a cell: how far the methylation entropy sits above the Hmin floor. The same ratio in every domain, from the same physics.
02
Minimal inputs — each instrument requires only what physics needs Astro-QAPE: one published gate error rate. Astro-SCAPE: three published chip specs (TDP, transistor count, clock frequency). Astro-GAPE: a mean methylation beta from the cell architecture class of concern — obtained from a blood draw using the appropriate assay. What that assay looks like in practice is one of the open questions prospective trials are designed to answer. Astro-QAPE and Astro-SCAPE inputs are publicly available. The Astro-GAPE input is personal.
03
The floor and ceiling are derived, not measured The framework derives the physics floor and the architecture ceiling (ISA floor, substrate inversion, Hmin) for each class from first principles. These values are not published by any manufacturer. They are what the physics says the limits are — independent of what any platform has achieved so far.
04
Three questions answered for every platform Where are you now within your operating range? How far to the floor — the physics minimum you can never cross? How far to the ceiling — the architecture limit you cannot pass without redesign? Everything else in the analysis (three-component decomposition, temperature sensitivity, trajectory, predictions) follows from these three numbers.
05
Zero adjustable parameters — validated against reality Nothing is fitted to observed data. The framework is derived from physics and applied to published inputs. Validated: Astro-SCAPE correctly called H100→H200 as a small step and Blackwell B200 as a large step before either shipped. Astro-GAPE correctly predicts the direction of epigenomic floor breach in 29 of 30 TCGA cancer types with no training data.

Foundational Physics Research

The commercial instruments are built on an active foundational physics research program — an independent derivation of why the physics works at every scale.

The Informational Actualization Model is an ongoing independent research program in foundational physics: cosmology, particle physics, quantum foundations, and the thermodynamics of information. The same principle that sets the efficiency floor for a quantum computing gate also governs the expansion rate of the universe — the thermodynamic cost of irreversible information production at every scale, from the Hubble horizon to the transistor junction.

The research is conducted openly. Every paper, every MCMC chain, every timestamped prediction is publicly available and independently verifiable. The cosmological constant has been derived to 0.07% agreement with observation. The baryon asymmetry has been derived to 0.36% — without nuclear physics input. Both from zero free parameters beyond Planck 2018.

View Research on GitHub
Cosmological Constant
0.07%
Derived: 1.38 × 10⁻¹²³ · Observed: 1.14 × 10⁻¹²³.
Closes 121.97 of 122 orders of magnitude. Zero free parameters.
Baryon Asymmetry
0.36%
Derived: η = 6.115 × 10⁻¹⁰ · Observed: 6.137 × 10⁻¹⁰.
No nuclear physics input. 18th converged MCMC chain.
Globular Cluster Age Tension
0.10σ
Predicted: 1.378 Gyr excess · Valcin+2025 finds 1.397 Gyr.
Timestamped March 2026. Not fitted to this result.
Axis of Evil
0.15σ
Predicted: AIAM = 0.0628 · Observed: 0.066 ± 0.021.
Timestamped March 2026. Not fitted to this result.
Particle Physics · Charged Lepton Mass Ratio
2/3
Koide formula derived from first principles. Q = (me + mμ + mτ) / (√me + √mμ + √mτ)2 = 2/3 emerges as a thermodynamic constraint from informational saturation.
📄 Read paper
Public Record — 38 Papers · 18 Chains · 9 Research Categories
I. The Cosmological Model
II. Thermodynamic Identity
III. Cosmological Constant
IV. Arrow of Time
V. Dark Sector Identification
VI. Black Holes & M–σ
VII. Quantum Foundations
VIII. Particle Physics
IX. Structure Formation
doi: 10.5281/zenodo.18702042
View Archive

Intelligence Reports

Dated, falsifiable analyses derived from published specifications and first-principles physics. A physics-derived floor and architecture ceiling for every class — built entirely from publicly available data. Astro-GAPE is open science. Astro-QAPE and Astro-SCAPE are commercial instruments.

Latest
Astro-QAPE
Quantum Hardware Intelligence Report — Issue 002
April 2026
The Race to Fault Tolerance · Substrate Inversion · Gate error decomposition · 13 platforms · Engineering Diagnostics
Download PDF
Astro-QAPE · Interactive Instrument
Quantum Intelligence Platform
Full instrument demo — all analysis panels, real published data. Read-only preview. Live access by arrangement.
Open Demo
Astro-QAPE · Foundational Physics Paper
Josephson Junction Quasiparticle Floor
PRL-format submission · 6 pages
Predicted minimum quasiparticle density xqp in superconducting Al/AlOx junctions: xqp,min = 2.0×10−7, derived with zero free parameters and consistent with nine independent published observations. The first-principles physics underneath what Astro-QAPE measures.
Download PDF
Astro-QAPE · Prior Issue
Quantum Hardware Intelligence Report — Issue 001
March 29, 2026
Gate fidelity · Material ceilings · Architecture classification · Competitive trajectory
Download PDF
Latest
Astro-SCAPE
Semiconductor Intelligence Report — Issue 002
April 2026
Apple M5 & NVIDIA B200 (new data) · Three-Component Efficiency Breakdown · Node Scaling Regime · Dennard Amplifier · Engineering Diagnostics · Four Dated Predictions (2 confirmed)
Download PDF
Astro-SCAPE · Interactive Demo
Semiconductor Intelligence Platform
Full instrument demo — all analysis modules, 22 chips in database. Read-only preview. Live access by arrangement.
Open Demo
Astro-SCAPE · Prior Issue
Semiconductor Intelligence Report — Issue 001
March 29, 2026
Process node analysis · Thermal scaling · Dennard transition · Competitive ranking
Download PDF
Latest
Astro-GAPE
Genomic Intelligence Report — Issue 002
April 2026 · 120 pages
Five-substrate cellular architecture framework · Hmin floor from first principles · 8 architecture-class cards · Three identified inversions · C1/C2/C3 decomposition · Clinical scenarios including reserve-depletion signal · Four priority dated predictions · Cellular-departure detection trajectory 2010-2030 · Zero training data
Download PDF
Astro-GAPE · Prior Issue
Genomic Intelligence Report — Issue 001
April 2026
Cell architecture thermodynamics · Hmin floor derivation · 29/30 cancer types · Neurodegeneration · Zero training data
Download PDF
Astro-GAPE · Interactive Demo
Cellular Intelligence Platform
Open science demo — enter a mean methylation beta and cell architecture class to compute an A-score against the physics-derived Hmin floor. Pre-clinical research tool. Not intended to diagnose any individual.
Open Demo

Continuously refined.
Continuously validated.

Every new published specification is an opportunity to test the framework against reality. Predictions are filed before data publishes. The record is public, dated, and cumulative.

Physics-grounded, not fitted
Every output the instruments produce is derived from the physics of the substrate and noise mechanism — not calibrated to historical performance data, not fitted to trends, and not adjusted after the fact. Zero adjustable parameters across all three instruments.
Architecture ceilings not published by anyone else
The physics floor and architecture ceiling for each class are IAMPerformance-derived — not published by AMD, Intel, Apple, NVIDIA, any quantum hardware manufacturer, or any cancer research institution. Derived from first principles and validated against published data. These are what physics says the limits are, independent of what any platform has achieved so far.
Semiconductor predictions confirmed before data published
Two Astro-SCAPE predictions confirmed before the hardware shipped: H100→H200 correctly called as a small step (7.6%). Blackwell B200 correctly called as a large step (51.6% confirmed, 6.8× larger than H100→H200). The framework separated a memory upgrade from an architectural advance using only published input specifications.
Cellular validation — zero training data
Astro-GAPE correctly predicts the direction of epigenomic floor breach in 29 of 30 TCGA cancer types (4,304 matched tumor-normal pairs) with zero free parameters and zero cancer training data. The threshold was derived from healthy-cell physics — not from cancer data. Pre-clinical research only. Not intended to diagnose, treat, or screen any individual. Prospective clinical validation has not been performed.

Get in touch.

Heath W. Mahaffey lived and worked in the Faroe Islands and Denmark for several years before returning to the United States. The signature on every IAMPerformance paper — Fri Forsker, the Nordic term for an independent researcher — is meant in the plain sense the Danes themselves use it. The tradition matters. Niels Bohr was its most famous exponent: a physicist who held himself accountable to the public record — to the work, to the data, to the openly stated prediction — rather than to any single authority above him. The framework you see on this site is built in the same posture. Every prediction is timestamped on a public GitHub repository before any cohort is scored. Every atlas calibration is sealed with a SHA-256 hash. Every claim is re-runnable from the source by any skeptical reader. The discipline is the descendant of that tradition. The work continues here.

IAMPerformance is the Inter-Domain Research Institute built on one discovery: every irreversible physical transition — from cosmic structure formation to qubit decoherence to transistor switching to cellular methylation maintenance — leaves the same dimensionless fingerprint. The Mahaffey Number measures how far any system sits above its physics floor and below its architecture ceiling. Three published instruments apply this to three domains: Astro-GAPE (genomic — open science), Astro-QAPE (quantum hardware intelligence), and Astro-SCAPE (semiconductor intelligence) — all derived from first principles, zero free parameters, independently verifiable. The underlying Architecture-class framework is proprietary (patent pending US 64/012,720 and US 64/014,568).

The philosophical and theological foundations that preceded this scientific work are in a published book — written before the derivations, before the instruments, before the data. Read the book →

Astro-GAPE research collaboration: Clinical validation partnerships are actively sought — especially for glioma grading, Alzheimer's pre-symptomatic signatures, BRCA carrier monitoring, and canine oncology. Astro-GAPE is pre-clinical research only and is not intended to diagnose, treat, or screen any individual.

Commercial instruments (Astro-QAPE/Astro-SCAPE): Live access by arrangement.

For inquiries about the framework, the APE engines, licensing, research collaboration, or the published predictions, reach out directly.

hmahaffeyges@gmail.com

Research, Astro-GAPE collaboration, and commercial (Astro-QAPE/Astro-SCAPE) — all inquiries

All commercial inquiries through legal counsel.
Patent Pending: US Provisional Applications 64/012,720 and 64/014,568.
Heath W. Mahaffey — Fri Forsker, Entiat, Washington.
Principal Researcher and Founder, IAMPerformance Inter-Domain Research Institute.

Intellectual Property
Astro-QAPE — Astro-Quantum Architecture-class Performance Engine
US Provisional 64/012,720 · Filed March 21, 2026
Full APE Platform — Universal Information Processing Systems
US Provisional 64/014,568 · Filed March 23, 2026
All inputs from publicly available published sources.
All framework outputs timestamped and verifiable.
IAMPerformance
iamperformance.net