Measure of species variability for a microbial taxonomy based on the relative resemblance

The concept of species currently in use in microbial taxonomy is based on sex because derives from that developed in zoology and botany. The absence of sex as the only system to reproduce does not allow to use the hybridization as the test to assess the conspecificity of microbial strains, forcing microbial taxonomists to use relative resemblance among strains as the only tool to define microbial species and to classify new microorganisms. Relative resemblance can be intuitively defined as the situation in which two strains of the same species must be more similar than each of them with a strain of any other species. Unfortunately, there are several algorithms to define the similarity between two strains, but none can be used to estimate the average distance between several members of the same species. This paper describes the problems inherent with the definition of species without hybridization tests and proposes two algorithms for a standard estimation of the overall variability of a species with the data obtained from a sample of strains. These measures will allow the non-subjective determination of the overall similarity among members of the same species using results from both phenotypic and molecular analyses. Both algorithms are based on a bootsrapping procedure implemented by RHO and SMA, two software applications freely available from the internet. These applications allow an easy evaluation of parameters such as the overall variability and levels of confidence associated with the sample of strains and the panel of characters under study.