SRA

Understanding the biology of native microbial communities is hindered by the lack of robust functional data for the microbes within these communities. Quantifying mRNA levels via transcriptomics to infer function has proven successful in these communities. However, this requires the ability to accurately predict protein levels, which are the primary functional units, from mRNA levels. While a positive correlation exists between mRNA and protein levels, for certain genes, mRNA is not a predictor of protein. To address this challenge, studies have quantified the protein-to-RNA (PTR) ratios of all genes, including those in which mRNA levels are not predictive of protein levels. These data enabled the calculation of RNA-to-protein (RTP) conversion factors for some of these genes that, when applied to mRNA levels, enhance the predictivity for protein levels. Despite the potential of RTP conversion factors, their calculation requires extensive datasets, which are costly and not available for most microbes. Here, we generated and analyzed comprehensive datasets from seven bacterial strains and one archaeon and identified orthologous genes in which mRNA was not predictive of protein but had consistent PTR ratios. Calculation and application of conversion factors for these genes improved protein prediction from mRNA, even when the conversion factors were derived from distantly-related bacteria. RTP conversion factors derived from bacteria also improved protein predictivity from mRNA in an archaeon, indicating that this approach is robust across domains of life. Ultimately, this approach improves protein prediction from mRNA without the need for paired transcriptomic/proteomic data from a microbe of interest.