Local-first computational biophysics for reproducible structural analysis Biomolecular structure treated as geometry under torsion — not as a black box.
Phidynamics is a local-first computational biophysics framework for structural analysis of biomolecules using explicit geometric observables derived from torsion, curvature, and constrained phase dynamics.
It replaces opaque inference with deterministic, inspectable, and reproducible geometric descriptors that can be computed directly from real structural data on consumer hardware.
Phidynamics provides a reproducible structural analysis pipeline for biomolecular systems.
Given a structural input (PDB or XYZ), it can:
- analyze structural geometry
- compute normalized geometric signatures
- compare structural similarity
- classify structural regimes
- simulate constrained phase dynamics
- generate integrated structural diagnostics
Phidynamics is designed for:
- local execution
- deterministic outputs
- geometric interpretability
- structural comparison
- falsifiable low-cost experimentation
It does not require:
- GPU clusters
- molecular dynamics supercomputing
- proprietary infrastructure
- machine learning training pipelines
All outputs are generated locally, deterministically, and reproducibly from structural input.
Clone the repository and create a virtual environment:
git clone https://github.com/vector1109/Phidynamics.git
cd Phidynamics
python -m venv venv
venv\Scripts\activate # Windows
# source venv/bin/activate # Linux / macOS
pip install -r requirements.txtpython main.py analizar 1BNA
python main.py analizar input/xyz/grafeno.xyzComputes structural signature, topology, and geometric classification.
python main.py comparar input/xyz/grafeno.xyz input/xyz/tetraedro.xyzComputes normalized structural similarity between two geometries.
python main.py buscar input/xyz/grafeno.xyz input/xyz/Finds nearest structural neighbors in a local dataset.
python main.py faseExecutes constrained torsion-phase evolution and reports absorption dynamics.
python main.py fase-sweep
python main.py fase-stats
python main.py fase-reportRuns multi-condition phase sweeps, summarizes stability, and reports dynamic regimes.
python main.py diagnosticar input/xyz/grafeno.xyzRuns the full MVP pipeline:
structure → geometry → phase → absorption → regime → diagnosis
Phidynamics produces explicit, inspectable observables such as:
- structural class
- dimensionality
- coordination
- cyclicity
- normalized torsion absorption
- phase regime
- integrated structural diagnosis
Example diagnostic output:
File: input/xyz/grafeno.xyz
Signature: graphene / hexagonal lattice
Dimensionality: 2D
Absorption: 1.935
Regime: stable
Memory: indeterminate
Conclusion: graphene / hexagonal lattice | stable regime | indeterminate memory
Phidynamics produces stable separation between normalized structural regimes under torsion-derived observables.
Globular proteins remain clustered in low-Δ bands, relaxed DNA-like states collapse near the lower structural regime, and canonical B-DNA occupies a reproducibly higher Δ band.
Phase absorption increases monotonically across simulation cycles, indicating stable convergence under iterative phase evolution.
The system exhibits ordered convergence rather than stochastic oscillation, supporting deterministic phase behavior under repeated local execution.
Repeated phase sweeps produce low final dispersion across independent runs.
Low sweep variance indicates that phase response is stable across perturbation runs and does not depend on isolated favorable initial conditions.
| Test | Result |
|---|---|
| Structural Δ separation | Stable |
| Phase convergence | Monotonic |
| Sweep robustness | High |
| Regime reproducibility | Stable |
| Local deterministic execution | Confirmed |
Phidynamics is an experimental computational biophysics framework.
It is intended for:
- structural modeling
- geometric comparison
- computational hypothesis generation
- exploratory biophysical analysis
It is not intended for:
- medical diagnosis
- clinical interpretation
- treatment decisions
- biomedical intervention
Current outputs should be interpreted as geometric observables, not direct biological causality.
Phidynamics investigates whether normalized torsion-derived geometry can act as a stable observable for distinguishing structural regimes in biological systems.
Instead of modeling biomolecular organization as a purely statistical or inferential problem, Phidynamics treats structure as constrained geometric response under torsional organization.
Its working hypothesis is that normalized geometric observables can reproducibly separate:
- globular protein regimes
- canonical B-DNA regimes
- relaxed low-tension conformations
This is currently treated as a computational biophysics hypothesis, not an established physical law.
The strongest current result is not the claim of a new physical law.
The strongest result is that Phidynamics produces a normalized torsion-derived observable that appears to separate structural regimes reproducibly under local execution.
Observed Δ ranges currently show stable separation between:
| Structural Class | Observed Δ Range |
|---|---|
| Globular proteins | ~0.04 – 0.09 |
| Relaxed DNA-like states | ~0.04 |
| Canonical B-DNA | ~0.17 – 0.18 |
This suggests Δ may function as a reproducible descriptor of structural regime under normalized geometric comparison.
That is the current central result of the framework.
Phidynamics does not attempt to compete with molecular dynamics by brute force.
Its contribution is different:
it proposes that low-cost geometric observables may capture biologically relevant structural regimes without requiring high-complexity simulation stacks.
This makes the framework:
- inspectable
- reproducible
- portable
- computationally cheap
- scientifically falsifiable
Current status: MVP stable baseline
Phidynamics is now:
- versioned
- reproducible
- locally executable
- structurally testable
- baseline-stable
Current baseline commit:
4ce778a
MVP estable: pipeline estructural + fase + diagnostico integrado
- Procrustes alignment baseline
- RMSD vs Δ benchmark
- batch multi-PDB CLI
- JSON export pipeline
- unit tests for Δ stability
- CI reproducibility validation
- A-DNA / Z-DNA extension
- harmonic parameter sweeps
- publishable validation report
- statistical significance analysis
Phidynamics uses a dual-license model.
All executable source code is released under:
GNU Affero General Public License v3.0 (AGPL-3.0)
This includes:
- source code
- execution scripts
- validation
- visualization
- reproducible tooling
Conceptual formulation, theoretical interpretation, and methodological framework are released under:
CC BY-NC-ND 4.0
This includes:
- theoretical framing
- conceptual methodology
- nomenclature
- explanatory formalism
- interpretive framework
The code is open for audit and extension. The conceptual framework is open for study and citation.
@software{phidynamics2026,
author = {Farias, Fabian Dario},
title = {Phidynamics: Local-first Computational Biophysics for Reproducible Structural Analysis},
year = {2026},
url = {https://github.com/vector1109/Phidynamics}
}Fabián Darío Farías fabianista / Vector Torsion SRL
Phidynamics does not claim consensus. It claims reproducibility.