Reference-Signature Projection Framework for Exercise Transcriptomic Directionality Based on the IAV-DI Reference Axis

Abstract

**Background:** Exercise is often described as medicine, but exercise and pharmacological interventions are not clinically equivalent. We developed a Reference Signature Projection Framework (RSPF) to evaluate external disease and exercise transcriptomic responses relative to a predefined disease-treatment axis. This proof-of-concept analysis was restricted to the IAV-DI axis defined by dimethyl itaconate (DI) treatment during influenza A virus (IAV) infection. **Methods:** The IAV-DI reference was derived from transcriptional differences between DI-treated and untreated human peripheral blood mononuclear cells in the context of IAV infection. Drug-UP and Drug-DOWN genes were defined by avg_log2FC, prioritized by random forest Mean Decrease Accuracy, and represented by the MDA Top 100 signature. The two gene sets were separately projected by GSEA onto complete ranked gene lists for GSE243629 influenza A (IA) versus healthy controls (HC), cardiopulmonary exercise testing (CPX), and Marathon post- versus pre-exercise comparisons. Robustness and interpretive boundaries were assessed using gene-level similarity, TopN sensitivity, random gene-set controls, and pathway-level NES patterns. **Results:** The MDA Top 100 comprised 24 Drug-UP and 76 Drug-DOWN genes. In GSE243629 IA versus HC, Drug-DOWN was significantly enriched toward the upregulated end (NES = 1.724, FDR = 4.00 × 10⁻⁸), whereas Drug-UP was not significant, indicating an opposite disease relationship relative to the IAV-DI axis supported mainly by Drug-DOWN. Neither gene set reached FDR < 0.05 in CPX or Marathon; consequently, exercise direction scores were treated as exploratory. Gene-level analysis also showed a weak overall CPX–IAV-DI relationship. Selected pathways had the same NES sign for IAV-DI and CPX and the opposite sign for Marathon, but none of the CPX pathways displayed in Figure 9 reached FDR < 0.05, limiting these findings to exploratory module-specific patterns. **Conclusion:** RSPF evaluates transcriptomic directionality relative to a predefined disease-treatment axis. In this IAV-DI analysis, the external influenza comparison showed a statistically supported opposite relationship, whereas neither CPX nor Marathon showed significant overall IAV-DI directional concordance. The findings are specific to this reference axis and do not establish a universal exercise–drug relationship, therapeutic equivalence, drug substitution, or shared pharmacological mechanisms.