Exposure to spaceflight alters gene expression in multiple tissues, with the liver being among the most studied and affected due to its central role in metabolism and immune regulation. While most transcriptomic analyses nowadays rely on differential expression methods, machine learning (ML) models offer alternative approaches to uncover changes in gene expression patterns. In this study, we analyzed mouse liver RNA-seq data from NASA’s GeneLab study OSD-379, which includes samples from mice that stayed on Earth and mice that were flown to the International Space Station. To identify genes whose expression could differentiate mice that underwent spaceflight from those that stayed on Earth, we trained three supervised classification models: Decision Tree, Logistic Regression and Random Forest. A hyperparameter optimization was implemented using Bayesian search with Optuna while assessing the interpretability using SHAP (Shaplet Additive Explanations). Random Forest achieved the highest predictive performance, with a macro-averaged F1-score of 86.9%, outperforming Logistic Regression (78.3%) and Decision Tree (54.4%). SHAP analysis highlights candidate genes whose expression patterns are strongly associated with spaceflight exposure. The results presented in this work demonstrate that optimized ML classifiers can distinguish transcriptomic profiles related to spaceflight with high accuracy and interpretability. Our work represents a scalable pipeline for future applications in space omics, with potential to accelerate the discovery of new molecular signatures responsive to spaceflight stressors.