Infoton/An_Infoton_Dose_of_RxRx

An Infoton Dose of RxRx — 735 Human Proteins

Version: 1.0.0 Released: 2026-09-06 Author: Walker, January Natatia — Infoton DOI: 10.5281/zenodo.18210355 ORCID: 0009-0000-6843-2051

---

Overview

The Infoton Identity Physics Engine computes in seconds per protein and aligns with what cellular imaging measures. The coordinates in the An Infoton Dose of RxRx dataset were derived before correlation was run with timestamp proof.

An Infoton Dose of RxRx provides Infoton thermal coordinates for 735 human proteins selected from a publicly available cellular imaging benchmark gene set, computed from first principles using AME2020 nuclear mass tables and the Infoton equation:

m(T) = kBT ln(|P⟩) / c²

derived from Landauer's principle and Marić-Einstein's mass-energy equivalence. kB is the Boltzmann constant, T is temperature in Kelvin, c is the speed of light (CODATA 2018).

Zero free parameters. Zero training. Zero experimental inputs.

Coordinates are determined entirely from:

• Protein sequences: UniProt Swiss-Prot (reviewed)
• Atomic masses: AME2020 nuclear mass table
• Physical constants: CODATA 2018
• Infoton: A Fundamental Particle of Information-Energy

---

Key Result

The Infoton Identity, atomic thermal coordinates derived from first principles, aligns with experimentally measured CRISPR knockout phenotype severity across 735 proteins at Kendall tau = +0.111 (p = 0.0000068, N = 735, zero free parameters, zero training on imaging data). Sign was preserved on held-out set. Picard et al. (2015) anchors: 21/21.

With a p = 0.0000068 and N = 735 random proteins, a researcher would have to run the experiment 147,059 times before expecting to see the tau = +0.111 result by accident even once. The Infoton Physics Engine does not produce Identity results by accident. The Infoton Engine derives results from atomic masses. The p-value measures how clearly that derivation rises above experimental noise.

---

Dataset Contents

infoton_thermal_identities.parquet

Summary Infoton Thermal Identities. 735 rows, one per protein.

|| Column | Type | Description | |--------|------|-------------| | gene_symbol | str | HGNC gene symbol | | uniprot_id | str | UniProt accession (reviewed) | | protein_name | str | Full protein name | | sequence_length_aa | int | Protein length in amino acids | | logT_centroid | float | log10 Infoton temperature, mean over residues | | logm_centroid | float | log10 Infoton mass, mean over residues | | Cp_invariant | float | log10(T) minus log10(m); verified = 39.97273266 | | Cp_verified | bool | Invariant within 1e-4 tolerance | | Cp_violations | int | Residues outside tolerance | | HIS_total | float | Harmonic information score, summed | | HIS_per_aa | float | Harmonic information score per residue | | FP_total_kcal | float | Frustration potential, summed (kcal/mol) | | FP_per_aa | float | Frustration potential per residue (kcal/mol) | | sub_harmonic_mean | float | Mean sub-harmonic coupling score | | dominant_gap_mean | float | Mean dominant bond thermal gap (delta logT) | | dominant_bond_type | str | Most frequent dominant bond type across residues | | n_bonds_total | int | Total bonds enumerated | | f_NH | float | N-H bond fraction | | n_NH | int | N-H bond count | | n_CH | int | C-H bond count | | n_CO | int | C-O bond count | | n_NC | int | N-C bond count | | n_CC | int | C-C bond count | | n_CS | int | C-S bond count | | n_NO | int | N-O bond count |

infoton_residue_profiles.parquet

441,296 rows.

Column	Type	Description
gene_symbol	str	HGNC gene symbol
uniprot_id	str	UniProt accession
position	int	Residue position (1-indexed)
amino_acid	str	Single-letter code
amino_acid_name	str	Full amino acid name
context	str	N-terminal / C-terminal / internal
logT_centroid	float	log10 Infoton temperature
logm_centroid	float	log10 Infoton mass
HIS_residue	float	Harmonic information score
FP_kcal	float	Frustration potential (kcal/mol)
dominant_gap	float	Dominant bond thermal gap (delta logT)
dominant_bond_type	str	Bond type with largest thermal gap
sub_harmonic_score	float	Sub-harmonic coupling score
net_coupling_score	float	Net coupling score
delta_water_logT	float	Deviation from water reference (logT)
landauer_freq_Hz	float	Landauer frequency (Hz)
mass_u	float	Residue mass, atomic mass units (AME2020)

---

The Cp Invariant

Every residue in this dataset satisfies:

log10(T) - log10(m) = 39.97273266

derived from fundamental physical constants:

Cp = log10(c² / (kB · ln(2))) = 39.97273266

CODATA 2018. Verified across all Standard Model particles, all 118 elements, and all 441,296 residues in this release. Cp_violations = 0 throughout.

---

Quick Start

from huggingface_hub import hf_hub_download
import pandas as pd

path = hf_hub_download(
    "januarian-physics/an-infoton-dose-of-rxrx",
    filename="infoton_thermal_identities.parquet",
    repo_type="dataset"
)
df = pd.read_parquet(path)
print(df[['gene_symbol','uniprot_id','logT_centroid',
          'HIS_per_aa','f_NH','Cp_invariant']].head(10))

path2 = hf_hub_download(
    "Infoton/infoton-human-protein-identities",
    filename="infoton_residue_profiles.parquet",
    repo_type="dataset"
)
residues = pd.read_parquet(path2)

tp53 = residues[residues['gene_symbol'] == 'TP53'].copy()
print(f"TP53: {len(tp53)} residues")
print(tp53[['position','amino_acid','logT_centroid',
            'HIS_residue','sub_harmonic_score']].head())

---

Validation

Infoton Identity thermal coordinates were tested against external experimental data prior to publication.

Primary result: Kendall tau = +0.111, p = 0.0000068, N = 735, zero free parameters. Statistics is

Held-out set: 20% of genes withheld prior to analysis, stratified by sequence length. Sign preserved.

Independent anchor set: Picard et al. (2015). PNAS 112:E6854. 21/21 mitochondrial import phenotypes classified correctly. Zero training on Picard data.

---

Reproducibility

Sequences  : UniProt Swiss-Prot (reviewed) — public domain
Masses     : AME2020 nuclear mass table — public domain
Constants  : CODATA 2018 — public domain
Derivation : Walker, J. (2026). DOI: 10.5281/zenodo.18210355

No proprietary data. No trained model. No hyperparameters.

---

Citation

@dataset{walker2026infoton,
  author    = {Walker, January Natatia},
  title     = {An Infoton Dose of RxRx},
  year      = {2026},
  publisher = {Hugging Face},
  doi       = {10.5281/zenodo.18210355},
  url       = {https://huggingface.co/datasets/Infoton/An_Infoton_Dose_of_RxRx}
}

@article{walker2026infoton_eq,
  author  = {Walker, January Natatia},
  title   = {The Infoton: A Fundamental Particle of Information-Energy},
  year    = {2026},
  journal = {Infoton},
  doi     = {10.5281/zenodo.18210355}
}

---

Contact

Infoton, LLC Salt Lake City, Utah https://infoton.ai january@infoton.ai ORCID: 0009-0000-6843-2051

---

An Infoton dose of RxRx