We swept every known diabetes variant
and found a universal endpoint
A computational disease cascade analysis across all 76 documented diabetes variants — Type 1, Type 2, MODY, neonatal, and syndromic — identified 5 novel findings, 8 universal terminal nodes shared by every variant, and 3 repurposable compounds including a T2D drug that may treat T1D.
The PHYSIM Platform: A deterministic computational physics platform. Governed by strict mathematical laws, the system maps the entire human genome across every known disease to compute biological certainty, not generative probability. AI serves only as our translator — the core analysis is reproducible, auditable, and deterministic.
This does not replace the laboratory — it de-risks before you get there. Instead of screening thousands of candidates blindly, the system narrows the search space to a focused set of computationally validated targets worth testing. Each finding on this page is a possible new discovery — a possible path toward helping patients — that deserves rigorous experimental validation.
Every diabetes variant ends at the same 8 nodes
76 genetically distinct diabetes variants — T1D, T2D, MODY, neonatal, syndromic — all converge on the same structural endpoint. The path varies. The destination does not. Over 4.7 trillion mathematical computations produced this result.
⚡ = Novel finding — the GCK → IL2RA structural bridge was identified computationally. No published model describes GCK as the pressure node that collapses the T1/T2 classification boundary. Also present in all 76 variants: PLG (fibrinolysis failure), FUS (neurological decay), DMD (muscle integrity).
5 pathways the literature hasn't connected — here are the 3 most significant
These findings emerged from the computational cascade analysis across all 76 diabetes variants. Each represents a potential therapeutic breakthrough with testable predictions.
GCK → IL2RA Bridge: The T1/T2 Boundary Is Artificial
GCK (Glucokinase) — the glucose sensor in beta-cells — sits at the structural boundary between the metabolic arm (T2D) and the autoimmune arm (T1D). When GCK fails, the pressure propagates directly into the IL2RA T-cell regulatory space.
This means Type 1 and Type 2 are not separate diseases — they are different entry points into the same structural collapse. The classification boundary is a measurement artifact.
NKX2-5 Ghost Anchor: A Cardiac Gene in the Diabetes Neighborhood
NKX2-5 is a cardiac master transcription factor with no canonical role in adult diabetes. The analysis found it embedded in 9 of 76 diabetes variants — exclusively in beta-cell transcription factor neighborhoods (PAX4, NEUROD1, HNF1B, PDX1).
In 2025, independent labs confirmed NKX2-5+ mesenchymal cells are required for normal islet formation. The engine detected this signal computationally before the mechanism was published.
PINK1 Bidirectional Axis: Parkinson's and Diabetes Share a Structural Node
PINK1 — canonically a Parkinson's gene — appears in 7 of 76 diabetes variants, specifically in the GCK, INS, IRS1, and SLC2A4 neighborhoods. It is not downstream of hyperglycemia — it is co-resident with the primary insulin machinery.
This implies PINK1 disruption simultaneously damages dopaminergic neurons AND beta-cell mitochondrial quality. PINK1 activators (currently in Parkinson's pipelines) may be triple-indication: neuroprotection + beta-cell rescue + insulin sensitivity.
What we produce. PHYSIM generates Computational Disease Analysis (CDA) reports — pre-clinical computational investigations that map disease pathways, identify repurposable FDA-approved compounds, and surface novel molecular candidates. Each report is a commissioned, disease-specific analysis built from the platform's complete structural knowledge of 20,000+ human genes and 10,000+ mapped diseases.
3 compounds the analysis confirmed — you already know them
These drugs were independently identified through the cascade analysis, validating the method against established clinical knowledge.
When these interventions matter in diabetes progression
The cascade follows a universal sequence regardless of diabetes subtype. Early intervention at upstream nodes prevents downstream terminal lock engagement.
76 distinct genetic entry points → metabolic or autoimmune initiation
T1D (autoimmune), T2D (metabolic), MODY (monogenic), neonatal, syndromic. Different doors, same building.
GCK dysfunction → impaired glucose sensing → metabolic pressure
Glucokinase is the gateway enzyme. Dorzagliatin (GCK activator) targets this stage directly.
⚡ GCK → IL2RA pressure transfer → T-cell regulatory breach
The structural bridge between metabolic and autoimmune arms. GCK activation may simultaneously suppress autoimmune onset — a dual-indication thesis.
CTLA4 checkpoint failure → unbounded T-cell assault on islet remnants
Autoimmune final collapse. Abatacept (CTLA4-Ig) is already in T1D clinical trials targeting this node.
⚡ PINK1 co-resident with GCK/INS/IRS1 → Parkinson's-diabetes shared vulnerability
PINK1 activators from Parkinson's pipelines may protect beta-cell mitochondria. Triple-indication potential.
8 terminal nodes engage → coagulopathy, neuropathy, cardiomyopathy, insulin blindness
All 76 variants arrive here. FGA/FGB/FGG (coagulation), PLG (fibrinolysis), INSR (resistance), FUS (neurological), DMD (muscle), CTLA4 (immune). The endpoint is fixed.
Commission a Computational Disease Analysis
We produce pre-clinical CDA reports that map your disease of interest across the full human genome — identifying novel pathways, repurposable compounds, and structural drug targets before you enter the lab. Tell us what you're working on.
Start a Conversation →