Using Statistical Properties of Short Subsequences in Microbial Identification