Computational analysis of the specificity of 16S rRNA-derived signature sequences for identifying food-related microbes

Abstract

The efficacy of oligonucleotide-based molecular diagnostics for the identification of food-related microbes is critically dependent on the sequences of the gene probes or amplification primers used. Computational specificity analysis was used to evaluate and compare the apparent specificity of ribosomal small subunit (SSU)-derived signature sequences from a wide range of food-related microbes. Nine signature sequences were obtained from the SSU sequences of each of 40 species, the apparent specificity of each signature sequence was determined by using the Ribosomal Database Project (RDP) microbial SSU dataset. Use of the RDP database made it possible to compare results from a phylogenetically heterogeneous group of microbes. At least one species-specific signature sequence was found for 39 of the 40 species. Signature sequences derived from the SSU V1 and V2 variable regions were the most specific. The apparent specificity of each of 77 SSU-based identification sequences obtained from the literature was also tested. Although a relatively large number of the published sequences were either non-specific or could not be matched to homologous sequences in the RDP database, most species-specific published sequences also originated from the V1 and V2 variable regions. These results show that computational specificity analysis and sequence databases can be used to assess and compare the specificity of SSU-derived signature sequences for any food-related microbe, but that the addition of new sequences to the databases necessitates periodic reassessment of specificity.