Space Mineral Atlas · NASA OSDR & LSDA Reanalysis

SpaceMineralAtlas

Crop-specific mineral responses to spaceflight, discovery of a 1000-fold unit error in OSD-655, dietary supply sufficiency vs overload safety, and the NASA LSDA data access landscape.

Richard Barker · all values computed from public NASA APIs · source code

Read the manuscript (PDF) Code & data (GitHub) NASA OSDR

Overview

Key findings and reproducibility benchmarks across the SpaceMineralAtlas pipeline.

60

Plant samples

OSD-745 lettuce (24) & OSD-655 Tokyo Bekana (36).

2,676

LSDA experiments

423 mineral & nutritional studies audited.

1000×

Unit error fixed

OSD-655 mislabelled mg/g vs true µg/g dry weight.

62

Open LSDA studies

213 request-gated • 145 never submitted.

0

Synthetic points

Protected by CI regression & synthetic data guards.

1 A 1000-Fold Unit Defect in Open NASA Data (OSD-655)

Curated data correction of five major macronutrient elements.

OSD-655 (Tokyo Bekana grown in Veg-04A/04B) labels five major macronutrient elements — calcium, potassium, magnesium, phosphorus, and sulfur — as mg/g dry weight in both its data files and its official archived data dictionary. The true values are µg/g dry weight.

View Mode:
Macronutrient Concentration Implied Dry Mass % Biological & Physical Assessment
Potassium (K) 29,430 mg/g 2,943.0% Physically Impossible (>100%)
Calcium (Ca) 15,820 mg/g 1,582.0% Physically Impossible (>100%)
Sulfur (S) 8,950 mg/g 895.0% Physically Impossible (>100%)
Phosphorus (P) 6,840 mg/g 684.0% Physically Impossible (>100%)
Magnesium (Mg) 3,110 mg/g 311.0% Physically Impossible (>100%)

Physical Impossibility

Stated units in OSD-655 imply plant tissues are composed of 2,943% potassium by dry mass, exceeding total physical sample mass by nearly 30-fold.

Any analysis pooling OSD-655 with OSD-745 or other elemental datasets without correcting this defect is wrong by a factor of 1000 on five of twelve elements. A full curation report with provenance has been deposited for the OSDR curation team: View OSDR Curation Report →

2 Crop Divergence & Centred Log-Ratio (CLR) Compositional Analysis

Apparent calcium and iron increases are dry-matter dilution artefacts.

Raw elemental concentrations are compositional: if flight tissue shifts in dry-matter composition, all elements shift together even if specific ion uptake is unchanged. Correcting for dry-matter dilution using a Centred Log-Ratio (CLR) transform reveals that apparent calcium and iron increases are dilution artefacts.

Select Crop:
Element Raw Fold Change Raw p (FDR) CLR Fold Change CLR p (FDR) Biological Status After CLR Correction

Key Biological Finding

Lettuce shows zero elements significantly altered by spaceflight on either raw concentration or CLR scale. Tokyo Bekana displays significant CLR increases in potassium (+15%), sodium (+80%), and sulfur (+28%), while copper, manganese, and zinc exhibit relative decreases. Apparent calcium and iron accumulations do not survive correction.

Figure 1: Fold changes before and after Centred Log-Ratio (CLR) correction
Figure 1: Flight/ground fold changes before and after CLR transformation. Click to enlarge.
Figure 2: Crop divergence across elemental profiles
Figure 2: Tokyo Bekana vs Lettuce elemental response divergence. Click to enlarge.

3 Machine Learning: No Flight Signature Transfers Across Crops

Cross-crop validation reveals the absence of a universal microgravity ionome response.

Leave-One-Mission-Out (LOMO)

5 folds partitioning VEG-01A, VEG-01B, VEG-03A, VEG-04A, VEG-04B.

0.608

+0.040 Margin over chance (0.568)

Modest signal within mission cohorts, tracking environmental differences.

Leave-One-Crop-Out (LOCO)

Train on Lettuce → Test on Tokyo Bekana (and vice versa).

0.486

-0.113 Below chance (0.600)

Below chance performance: Confirms no shared signature transfers between crops.

Random Forest Feature Importance Ranking

Permutation importance confirms that only Potassium, Sodium, and Sulfur contribute positive predictive signal — matching the exact three elements that survived CLR compositional correction.

Figure 6: Model performance against baselines
Figure 6: Machine learning classification performance across cross-validation regimes. Click to enlarge.

4 Interactive Dietary Supply & Overload Safety Calculator

Real-time nutritional modeling against NASA JSC-67378 Rev 1 astronaut dietary standards.

Harvest Mass: 100 g
Figure 5: Serving contribution against requirements
Figure 5: Nutritional contribution of 100g fresh produce vs NASA JSC-67378 standards. Click to enlarge.

5 Inspiration4 Crew Clinical Dynamics & CDC NHANES Benchmarks

Longitudinal serum chemistry benchmarked against age- and sex-stratified US population percentiles.

Jared Isaacman (C001)
Commander • 38M
Matched: Males 30–49 (N=669)
Hayley Arceneaux (C002)
Medical Officer • 29F
Matched: Females 18–29 (N=494)
Chris Sembroski (C003)
Mission Specialist • 41M
Matched: Males 30–49 (N=669)
Sian Proctor (C004)
Mission Pilot • 51F
Matched: Females 50–69 (N=931)
Select Analyte:

Methodological Caveat

4 subjects, a 3-day mission duration, and 0 in-flight biosamples. These trajectories describe clinical observations but cannot establish causal microgravity mechanisms without long-duration in-flight sampling.

Inspiration4 vs NHANES Comparison Grid
Figure 4b: Inspiration4 pre/post percentiles vs NHANES population norms. Click to enlarge.
Figure 4: Astronaut serum trajectories
Figure 4a: Longitudinal clinical chemistry across L-92 to R+90. Click to enlarge.

6 Interactive Plant Sample Elemental Dataset (N=60)

Complete harmonised elemental ICP-OES measurements across OSD-745 and OSD-655 (µg/g dry weight).

60 samples
Sample ID Crop Mission Condition K (µg/g) Ca (µg/g) Mg (µg/g) Fe (µg/g) Mn (µg/g) Na (µg/g)

7 NASA LSDA Data Access Landscape Explorer

Audit of 2,676 LSDA experiments categorised by accessibility tiers.

62

Openly Available (14.7%)

Direct download without gating.

213

Request Required (50.4%)

Institutional IRB / Form approval needed.

145

Never Submitted (34.3%)

Completed studies without archived data.

High-Value Target: SMO-016E (Nutritional Status Assessment)

The premier dataset for linking spaceflight diet directly to crew serum mineral dynamics — 27 astronauts across ~165-day missions with paired in-flight dietary intake and serum chemistry — is request-gated. A complete, turnkey LSDA data request application has been prepared in the repository: Explore LSDA Request Package →

ID Study Title Principal Investigator Subject / Species Access Status
Loading catalog...
Figure 7: LSDA mineral data landscape
Figure 7: Full breakdown of NASA LSDA experiment catalog. Click to enlarge.

8 Reproducibility, FAIR Data & Research Downloads

Automated verification, zero synthetic data guarantee, and dataset repositories.

✓ Zero Synthetic Data Policy

Automated CI tests fail the build if numpy.random, synthetic data generation, or unvalidated fallbacks are introduced.

✓ LaTeX Macro Injection Audit

All 179 statistics in the manuscript are injected directly from analysis JSON outputs via LaTeX macros, eliminating manual typing errors.

Direct Research Downloads

📊
Plant ICP-OES Data CSV • Harmonised µg/g DW
🧑‍🚀
Astronaut × NHANES Data CSV • Percentile Ranks
📑
LSDA Mineral Catalog CSV • 423 Annotated Studies
📋
OSDR Curation Report Markdown • OSD-655 Defect
📕
Complete Manuscript PDF • Publication Format
💻
Source Code & Tests GitHub • MIT License

Cite This Research

@article{Barker2026_SpaceMineralAtlas,
  author    = {Barker, Richard},
  title     = {SpaceMineralAtlas: Crop-Specific Mineral Responses to Spaceflight, a Unit-Labelling Defect in OSD-655, and a Survey of Mineral and Dietary Data Availability Across the NASA Life Sciences Data Archive},
  year      = {2026},
  publisher = {GitHub / Zenodo},
  url       = {https://github.com/dr-richard-barker/SpaceMineralAtlas}
}