Regardless of computer hardware and software improvements, DNA sequence analysis is still very computationally complex and time-consuming. Therefore, novel approaches to tackle this challenge are more than needed. In our study, we proposed a k-mers feature extractionmethod, which can identify informative k-mers. Such search is vital, among others, for microbial source tracking, where we need to define primers for an in-silico PCR. The results demonstrate that the proposed method is able to produce better results than the compared n-grams approach, resulting in better clustering outcomes.