tRNA is a molecular thermometer of species' optimal growth temperature

Non coding RNAs have the potential to serve as a molecular thermometer due to their rapid and specific responses to heat. To investigate the relationship between non coding RNAs with optimal growth temperature (OGT), we analyzed tRNA and rRNA sequences from 1957 species across archaea, bacteria, and eukaryotes. Our results revealed that tRNA composition correlates more strongly with OGT than rRNA, particularly in archaea. Trinucleotide analysis showed thermophiles tend to have tRNA enriched with charged and hydrophobic anticodons, potentially contributing to codon bias and protein thermal stability. UMAP dimensionality reduction uncovered distinct compositional clustering patterns aligning with species' thermal preferences. Machine learning models trained on RNA features accurately classified species into thermophilic prokaryotes, mesophilic archaea, mesophilic bacteria, and eukaryotes, with tRNA features contributing most to performance. Feature importance analysis further emphasized the dominant role of tRNA in thermal adaptation, suggesting a thermosensory function. These findings highlight the evolutionary and adaptive significance of tRNA and rRNA sequences in shaping organisms' thermal tolerance and offer new perspectives on the link between genetic information and environmental adaptation.