Development of Advanced Analytical Tools Integrating External Datasets with MoTrPAC Data:Utilizing Statistical, Machine Learning, and AI Approaches to Uncover Molecular Mechanisms of Physical Activity - PROJECT SUMMARY/ABSTRACT The molecular mechanisms driving physiological adaptations to exercise remain insufficiently understood, despite their critical role in promoting human health. The Molecular Transducers of Physical Activity Consortium (MoTrPAC) has generated an extensive multi-tissue transcriptomic dataset from Fischer rats undergoing endurance training, offering an unprecedented opportunity to investigate systemic responses to physical activity (PA). However, current analytical methods fall short in capturing the complex regulatory programs that operate across tissues, timepoints, and species, limiting translational insights into human biology. This project introduces a transformative computational framework leveraging GeneCompass, a state-of-the-art foundational model pretrained on over 100 million human and mouse single-cell transcriptomes. GeneCompass integrates diverse biological knowledge, including promoter sequences, transcription factor binding sites, and gene regulatory networks, to create biologically informed embeddings that enable precise cross-species comparisons. We will extend and fine-tune GeneCompass to analyze MoTrPAC’s rat RNA sequencing data, bridging molecular adaptations in rats to human biology. By aligning orthologous genes across species, this approach will uncover conserved transcriptional regulators and pathways, revealing shared mechanisms of exercise-induced adaptations. Our project focuses on three key aims: adapting GeneCompass for multi-species transcriptomic analysis, performing high- resolution tissue- and timepoint-specific regulatory analyses of PA-induced adaptations, and translating findings from rat models to human health using cross-species comparisons and in silico perturbation experiments. These efforts will produce innovative computational tools, dynamic molecular models, and actionable regulatory maps that enhance the value of MoTrPAC data. By providing a systematic framework for understanding the molecular underpinnings of exercise, this research will inform precision medicine and support the development of evidence-based strategies to improve human health through physical activity.
