Integrating co-expression network analysis and machine learning to reveal the regulatory landscape of <i>GPD</i> genes in <i>Chlamydomonas reinhardtii</i> under salinity stress

Background: genes are known to be salt-responsive with high sequence identity and shared metabolic roles, the regulatory mechanisms of their differential expression have remained unclear. Here, we leverage publicly available time-course RNA-seq data to dissect the transcriptional regulation of these genes, using an integrative network analysis and machine-learning framework. Methods: -regulatory elements. Module robustness was independently evaluated using a Random Forest classifier to assess the separability of gene-to-module assignments. Results: -regulatory motifs. Random Forest validation confirmed strong module coherence, providing independent support for the inferred regulatory programs. Conclusions: are embedded in temporally and regulatory distinct transcriptional networks during salinity stress. By integrating co-expression analysis with supervised machine learning for internal module validation, this study introduces a robust strategy to refine regulatory inference from transcriptomic data. More broadly, it highlights the value of data reuse and integrative analysis for uncovering regulatory divergence among genes, with potential implications for functional optimization in microalgae and related biological systems.