Methods & Code
Open methods and reproducibility
The Institute's work favors methods that can be inspected, re-run, and challenged: derivations stated explicitly, code deposited publicly, and data released where it can be shared responsibly.
Computational methods
Recurring methodological tools across the programs.
Monte Carlo validation
Large-scale synthetic-cohort validation of clinical decision tools, reported with precision bounds rather than point claims.
Survival & competing risks
Structural-censoring corrections for incidence estimation, including survival-biased effective MDRI and joint bias factors.
Signal processing
Spatial-autocorrelation and correlation-length measures for neurometabolic biomarker discovery.
Algorithmic bias epidemiology
Quantifying synergistic, multiplicative discrimination that is invisible to single-feature fairness checks.
Within-host kinetics
Deriving population-level intervention bounds from mechanistic within-host viral dynamics.
Model reporting integrity
Structured, TRIPOD+AI-aligned assessment of whether prediction models are reported and validated adequately.
Reproducibility practices
- Persistent IDs Every output carries a DOI; code and analysis are archived on Zenodo.
- Public code Analysis repositories are released under the Nyx-Dynamics GitHub organization.
- Open data Datasets are released where they can be shared responsibly, with provenance documented.
- Explicit derivations Mathematical results are stated from first principles rather than empirical convention.
- Versioned preprints Preprints are versioned; superseding versions are linked to their predecessors.
Open data
Phase-specific HIV neurometabolic MRS dataset
The first publicly available consolidated dataset of phase-specific HIV neurometabolic magnetic resonance spectroscopy data — 49 observations across 7 independent human cohorts spanning 1985–2021 — released as an independent contribution to the HAND neuroimaging field alongside the Noise Correlation Length preprint.
Code repositories
Public repositories under the Nyx-Dynamics organization.
- Prevention Theorem github.com/Nyx-Dynamics/Prevention-Theorem
- LAI-PrEP tool github.com/Nyx-Dynamics/lai-prep-bridge-tool-pub
- Algorithmic bias github.com/Nyx-Dynamics/algorithmic-bias-epidemiology-academic
- AI Readiness Framework github.com/Nyx-Dynamics/ai-readiness-framework
- Organization github.com/Nyx-Dynamics