BackgroundThe environment has been playing an instrumental role in shaping and maintaining the morphological, physiological and biochemical diversities of prokaryotes. It has been debatable whether the whole-genome Guanine-Cytosine (GC) content levels of prokaryotic organisms are correlated with their optimal growth temperatures. Since the GC content is variable within a genome, we here focus on the correlation between the genic GC content levels and the temperature range conditions of prokaryotic organisms.ResultsThe GC content levels in the coding regions of four genes were consistently identified as correlated with the temperature range condition when the association analysis was applied to (i) the 722 mesophilic and 93 thermophilic/hyperthermophilic organisms regardless of their phylogeny, oxygen requirement, salinity, or habitat conditions, and (ii) partial lists of organisms when organisms with certain phylogeny, oxygen requirement, salinity or habitat conditions were excluded. These four genes are K01251 (adenosylhomocysteinase), K03724 (DNA repair and recombination proteins), K07588 (LAO/AO transport system kinase), and K09122 (hypothetical protein).To further validate the identified correlation relationships, we examined to what extent the temperature range condition of an organism can be predicted based on the GC content levels in the coding regions of the selected genes. The 84.52% accuracy for the complete genomes, the 84.09% accuracy for the in-progress genomes, and 82.70% accuracy for the metagenomes, especially when being compared to the 50% accuracy rendered by random guessing, suggested that the temperature range condition of a prokaryotic organism can generally be predicted based on the GC content levels of the selected genomic regions.ConclusionsThe results rendered by various statistical tests and prediction tests indicated that the GC content levels of the coding/non-coding regions of certain genes are highly likely to be correlated with the temperature range conditions of prokaryotic organisms. Therefore, it is promising to carry out “reverse ecology” and to complete the ecological characterizations of prokaryotic organisms, i.e., to infer their temperature range conditions based on the GC content levels of certain genomic regions.