Mining K-mers of Various Lengths in Biological Sequences

Counting the occurrence frequency of each k-mer in a biological sequence is an important step in many bioinformatics applications. However, most k-mer counting algorithms rely on a given k to produce single-length k-mers, which is inefficient for sequence analysis for different k. Moreover, existing k-mer counters focus more on DNA sequences and less on protein ones. In practice, the analysis of k-mers in protein sequences can provide substantial biological insights in structure, function and evolution. To this end, an efficient algorithm, called VLmer (Various Length k-mer mining), is proposed to mine k-mers of various lengths termed vl-mers via inverted-index technique, which is orders of magnitude faster than the conventional forward-index method. Moreover, to the best of our knowledge, VLmer is the first able to mine k-mers of various lengths in both DNA and protein sequences.