Defense, jasmonate, ethylene, MAPK, WRKY, redox and ubiquitination. All 12 pass FDR < 0.01 in the root limma analysis.
AI × DIGITAL TWINS × SPACE BIOLOGY
A root defense program.
Reproducible, but context-dependent.
PlantCellTwin exposes the evidence behind a 12-gene Arabidopsis root defense-regulatory panel discovered in NASA GLDS-7 and tested against independent OSDR studies.
MAIN FINDINGThe GLDS-7 root response is not a universal spaceflight signature: tissue zone, gravity dose and experimental context materially change its direction and strength.
LIVE EVIDENCE / 01
Root-defense digital twin
SIMULATION ACTIVET+ 00:00:00
351root loci
FDR < 0.05
FDR < 0.05
ρ = 0.996R × Python
effect concordance
effect concordance
12 genesdefense-regulatory
root panel
root panel
7/11zone-structured
in GLDS-208
in GLDS-208
12/12positive gravity trends
in OSD-251
in OSD-251
TOP PROBE EFFECTS GLDS-7 FLIGHT–GROUND TWIN PROJECTION
DIGITAL-TWIN SCENARIOS 0 records
| # | Tissue | Condition | Exposure | Light | Index | Leading signal |
|---|---|---|---|---|---|---|
| Record a scenario to begin the comparative survey. | ||||||
EXTERNAL NASA VALIDATION GLDS-17 · 37 · 38 · 44 · 121
—replicated comparisons
—partial evidence
20contrasts tested
Candidate genes at p ≤ 0.01 were compared across studies. “Replicated” requires significant overlap and directional agreement after BH correction across the 20 comparisons.
| GLDS-7 tissue | External study | Shared genes | Same direction | Overlap FDR | Direction FDR | Conclusion |
|---|---|---|---|---|---|---|
| Loading cross-validation… | ||||||
SCIENTIFIC LAYER V2 MODERATED VARIANCE · GO · TRANSFER
—GO processes at FDR < 0.05
—predictive transfers
—rank-consistent contrasts
22,591loci evaluated per tissue
Tissue-specific linear models with empirical-Bayes variance moderation. GO enrichment uses biological processes; transfer analysis tests whether the GLDS-7 effect predicts directions observed in other NASA studies.
Strongest biological processes
Loading…
Out-of-study test
Loading…
TESTE DA DESCOBERTA ASSINATURA RADICULAR DE DEFESA
—
replicated pathway–study pairs
A pathway is considered replicated when it has FDR < 0.05 and at least two overlapping genes in another NASA dataset.
Loading evidence from NASA files…
RAW-FILE CONFIRMATION 36 CEL · RMA · PLATE BLOCKING
—root loci at FDR < 0.05
—RMA-confirmed pathways
—pathways positive on all three plates
—root RMA × MAS5 correlation
CONFIRMATORY COMPUTATIONAL RESULT
The root-defense signal survived complete replacement of preprocessing, control for plates 2, 4 and 6, and multiple-testing correction. This reduces the possibility that the discovery is specific to the MAS5 matrix.
INDEPENDENT REPRODUCTION R 4.4.3 · BIOCONDUCTOR · AFFY + LIMMA
—root loci at FDR < 0.05
—R × Python effect correlation
—root top-300 overlap
—same direction within overlap
STRONG COMPUTATIONAL REPRODUCIBILITY
The 36 CEL files were reprocessed with affy::rma and tested with limma::eBayes, retaining tissue and plate in the model. This agreement confirms computational stability; biological validation still requires a prospective experiment.
RESEARCH LOGIC / 02
From telemetry to a testable hypothesis
Environmental telemetry
Temperature, RH, CO₂ and radiation streams from selected OSDR experiments.
Time-series AI
Preprocessing, forecasting, anomaly detection and uncertainty estimation.
Virtual cell response
An interpretable visual layer maps model outputs to explicit cellular-state hypotheses.
Biological validation
Future comparison with omics and phenotype data to test—not assume—associations.
“Can an AI-assisted 3D Digital Twin reproduce, detect and explain anomalous environmental telemetry patterns—and support testable hypotheses about cellular response—in space-biology experiments?”
SCIENTIFIC BASIS / 03
Evidence behind the model
The demonstrator connects established space-biology findings with an explicitly conceptual response model. Each source below identifies what can be grounded now and what still requires validation with selected OSDR datasets.
ENVIRONMENTMicrogravity · radiation · habitat conditionsObserved exposure and mission metadata
→CELLULAR PATHWAYSMitochondrial stress · oxidative response · DNA regulationMulti-omics evidence and testable mechanisms
→DIGITAL TWINState trajectory · anomaly · uncertaintyModel prediction, never a clinical conclusion
PLANT-CELL TRACK
Model organism: Arabidopsis thaliana
The plant twin links spaceflight conditions to chloroplast redox stress, cell-wall remodeling, root-cell response, photosynthetic state and tissue-specific gene-expression hypotheses.
CHLOROPLAST / ROSCELL WALLROOT GROWTHLIGHT RESPONSEION TRANSPORTHSP101 / COR78
Fifteen challenges for generative AI applications to cell biology
Frames the requirements for generative models that move beyond molecular prediction toward cellular and multicellular behavior, including biological priors, perturbation response, validation and interpretability.
SUPPORTS Generative cell-state trajectories, explainability and validation design.
Read article ↗GeneLab: Omics database for spaceflight experiments
Describes NASA GeneLab as an open platform for spaceflight omics, curated datasets and experiment metadata—the data foundation for replacing synthetic inputs with traceable biological evidence.
SUPPORTS Dataset provenance, omics integration and reproducible pipelines.
Read article ↗The NASA Twins Study: A multidimensional analysis of a year-long human spaceflight
Integrates physiological and multi-omics measurements across a long-duration mission, demonstrating time-dependent biological change, individual variability and post-flight recovery.
SUPPORTS Longitudinal trajectories, baseline comparison and recovery states.
Read article ↗Comprehensive multi-omics analysis reveals mitochondrial stress as a central biological hub for spaceflight impact
Identifies mitochondrial dysregulation and associated immune, metabolic and oxidative-stress signals across astronaut and model-organism datasets.
SUPPORTS Oxidative-stress state, mitochondrial response and pathway-level alerts.
Read article ↗Space Omics and Medical Atlas (SOMA) across orbits
Brings together molecular, cellular, physiological and phenotypic spaceflight resources, supporting cross-mission comparisons and more complete biological context.
SUPPORTS Multi-scale validation, cross-study comparison and future model expansion.
Explore SOMA ↗NASA Open Science Data Repository
Provides standardized GeneLab multi-omics data and mission context for experiments affected by microgravity, radiation and other space-environment factors.
NEXT STEP Select one cell type, one experiment and matched control before calibrating the model.
Open OSDR ↗Meta-analysis of the spaceflight and microgravity response of the Arabidopsis plant transcriptome
Reanalyzes 15 NASA GeneLab experiments through common pipelines, reinforcing cell-wall and oxidative-stress responses while identifying light, hardware and assay method as important confounders.
SUPPORTS Plant markers, metadata controls and robust cross-experiment comparison.
Read article ↗Arabidopsis ecotypes reveal oxidative-stress signatures in spaceflight
Shows that genetic backgrounds differ in transcriptional response while retaining signatures associated with oxidative stress.
SUPPORTS Chloroplast ROS, genotype selection and uncertainty between ecotypes.
Find article ↗Arabidopsis spaceflight transcriptome by tissue
Compares roots, hypocotyls and leaves from replicated ISS experiments, revealing tissue-specific responses that whole-plant averages can hide.
SUPPORTS Root-cell mode, tissue selector and matched spaceflight-versus-ground controls.
Open GLDS-7 ↗MODEL BOUNDARY
Sample counts, effect sizes, moderated tests, GO enrichment and cross-study transfer come from reanalysis of public NASA/GEO data. Raw CEL normalization has now been reproduced independently in R/Bioconductor and Python. The exposure and light controls remain model projections—not new NASA measurements. Biological confirmation still requires preregistration and a prospective experiment.
PROJECT NOTE / 04
Designed for collaboration, not overclaiming.
This demonstrator separates NASA-derived measurements from model projections. Its empirical layers now combine tissue-aware differential expression, GO processes and cross-dataset transfer across five additional GeneLab studies. Recorded twin scenarios can be exported as a reproducible hypothesis ledger.
NASA GLDS-736 SAMPLES22,591 LOCIEXTERNAL VALIDATION