Vertebrate endurance capacity is a phenotype with considerable genetic heterogeneity. RNA-Seq technologies are an ideal tool to investigate the involved genes and processes, but several challenges exist when the phenotype of interest has a complex genetic background. Difficulties manifest at the level of results interpretation because commonly used statistical methods are designed to identify strongly associated genes. If an observed phenotype can be achieved though multiple distinct genetic mechanisms then typical gene-centric methods come with the attached risk that signal may be lost or misconstrued. Gene set analysis (GSA) methods are now widely accepted as a means to address some of the shortcomings of gene-by-gene analysis methods. We carry out both gene level and gene set level analyses on Xenopus tropicalis to identify the genetic factors that contribute to endurance heterogeneity. A typical workflow might consider gene level and pathway level analyses, but in this work we propose an additional focus at the intermediate level of functional modules. We generate functional modules for GSA testing in order to be explicit in how ontology information is used with respect to the functional genomics of Xenopus. Additionally, we make use of multiple assemblies to corroborate implicated genes and processes. We identified 42 core genes, 10 functional modules, and 14 pathways based on gene expression differences between endurant and non-endurant frogs. The majority of the genes and processes are readily associated with muscle contraction or catabolism. A substantial number of these genes are involved in lipid metabolic processes, suggesting an important role in frog endurance heterogeneity. Unsurprisingly, many of the gene expression differences between endurant and non-endurant frogs can be distilled down to the capacity to utilize substrate for energy, but at the individual level frogs appear to make use of diverse machinery to achieve these differences.