K2 and K2*: efficient alignment-free sequence similarity measurement based on Kendall statistics

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Traditional sequence alignment methods are computationally intensive for large datasets.
• Alignment-free methods offer a faster alternative for sequence comparison.

Purpose of the Study:

• To introduce K2 and K2*, novel alignment-free sequence comparison methods.
• To evaluate the performance and speed of K2 and K2* against existing methods.

Main Methods:

• K2: A non-parametric alignment-free method based on Kendall statistics.
• K2*: An improved version of K2 with automatic parameter selection.
• Comparative analysis with state-of-the-art alignment-free methods.

Main Results:

• K2 and K2* demonstrate competitive performance in phylogenetic tree generation and functional regulatory sequence evaluation.
• K2 and K2* significantly outperform other methods in speed and accuracy, especially for large datasets.
• K2* automatically determines optimal algorithmic parameters, enhancing usability.

Conclusions:

• K2 and K2* represent a significant advancement in alignment-free sequence comparison.
• The proposed methods offer a faster and more efficient alternative for analyzing large biological sequence datasets.
• The K2 and K2* R package is freely available for open access.