Interactive Data Lab
NEWRun custom Python analyses on Neurodiagnoses data
Secure, isolated execution environment with access to our Knowledge Base
FOR RESEARCH USE ONLY - NOT A MEDICAL DEVICE
This analysis platform is for research purposes only and not validated for clinical diagnostic use. Results are intended for scientific investigation and should not be used for medical decision-making.
How it works
1. Upload your CSV or Excel file with patient data below
2. Use AI Co-pilot to ask questions or write Python code manually
3. Code runs in a secure, isolated container with pandas, numpy, matplotlib, and seaborn pre-installed
📁 Data Source
Upload a CSV or Excel file to analyze
🧪 Reproducible Scientific Experiments
Test the Active Risk Hypothesis: Does combining genetic risk (PRS) with causal validation scores improve survival prediction beyond genetics alone?
This experiment trains two Cox proportional hazards models on a synthetic cohort and compares their predictive power.
Experiment Results
Static Model (PRS Only)
Active Model (PRS + Causal)
Cohort: patients analyzed
Clinical Interpretation
🧠 Spatial Concordance Hypothesis (Axis 2→3)
Test if molecular markers (Axis 2) preferentially express in atrophied brain regions (Axis 3). This validates the mechanistic coherence of our tridimensional framework.
Uses spatial validation logic to compute concordance scores and t-test for statistical significance.
Spatial Concordance Results
Concordance Score
Mean score (0-1)
Statistical Test
Atrophy Pattern Analysis
Relevant Regions (predicted by Axis 2):
Mean Z-score
Irrelevant Regions (control):
Mean Z-score
Scientific Interpretation
🧬 Replicación SOTA: Bachmann et al. (Preclinical Screening)
Epic 44Hypothesis: Plasma biomarker fusion (Aβ42/40 + p-tau217) outperforms single biomarkers for predicting MCI conversion in cognitively healthy (CU) individuals.
Replicates Bachmann et al. finding using NACC CU cohort with longitudinal follow-up (CN→MCI). Compares 3 Logistic Regression models: Aβ only, Tau only, and Fusion (A+T).
Preclinical Screening Results
Cohort Summary (CU Patients)
Total Analyzed:
CN→MCI Converters:
Conversion Rate:
Model 1: Aβ42/40 Only
Model 2: p-tau217 Only
Model 3: Fusion (A+T)
Improvements:
Clinical Interpretation
🔬 Retrospective Pharmacology Lab (Axis 0c)
Epic 52Full-Stack Industrialization: Run parameterized case-control studies on NACC data to evaluate the effect of any drug on cognitive progression (MMSE slope).
Uses Axis 0c (Pharmacology) architecture to identify patients on target medications, match controls by demographics/genetics, and compare progression rates. Evolved from Epic 50.
🔬 Retrospective Pharmacology Lab
Configure and run case-control studies using the Axis 0c (Pharmacology) architecture. Compare progression rates between patients on specific medications vs matched controls.
🗺️ Spatial Mechanism Validator (Axis 2 ↔ Axis 3)
Epic 53 v2Topology & Hallmark Validation: Query the NeuroPro tissue atlas (23K protein-region associations) to validate mechanism hypotheses from literature.
Maps biomarker candidates (Eje 2) to brain expression topology (Eje 3a) and neuropathological hallmarks (NFT, Plaque, CAA). Re-uses Epic 22 (SpatialValidator) infrastructure.
🗺️ Spatial Mechanism Validator
Query the NeuroPro Tissue Atlas (23K protein-region associations) to validate mechanism hypotheses. Maps biomarker candidates to brain expression topology and neuropathological hallmarks (NFT, Amyloid, CAA).
🧠 Spatiotemporal Atrophy Propagation Modeling
Predict personalized trajectories of cortical atrophy propagation using graph-based dynamical modeling (Li et al., arXiv:2511.08847)
🎯 Causal Fusion Hypothesis (Axes 1+2+3a→3b) - REAL DATA
Test if combining genetics (Axis 1), biomarkers (Axis 2), and imaging (Axis 3a) predicts clinical dementia (Axis 3b) better than each axis alone. This validates the core premise of the tridimensional framework.
Uses real NACC data (204K patients). Trains 3 classification models and compares AUC scores.
Causal Fusion Results
Model 1: Genetics
AUC score
Model 2: Biomarkers+Imaging
AUC score
Model 3: Fusion (All Axes)
AUC score
Model Comparison
Fusion vs Genetics:
Fusion vs Molecular:
Scientific Interpretation
📊 Validation Results: Genetics (Axis 1) → Molecular Biomarkers (Axis 2)
Pre-loaded analysis from our validation study (SEA-AD cohort, 84 patients)
Loading causal validation data...
🤖 Analysis Co-pilot
Ask questions in natural language