Simrank: Rapid and sensitive general-purpose k-mer search tool

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Terabyte-scale string-encoded data collections are emerging from large-scale biological projects like the Human Microbiome Project.
• K-mer matching strategies are crucial for efficient data similarity searches within and across projects.
• Existing software benefits from embedded k-mer searches, but a versatile, stand-alone tool was lacking.

Purpose of the Study:

• To introduce Simrank, a novel stand-alone utility for rapid identification of similar strings within large datasets.
• To provide a flexible, open-source k-mer tool for diverse applications.

Main Methods:

• Development of the Simrank utility for k-mer based string similarity searching.
• Performance benchmarking of Simrank against existing tools using various datasets (DNA, RNA, protein, human language).

Main Results:

• Simrank demonstrates significant speed improvements, ranging from 10X to 928X faster than related tools.
• The tool's performance advantage is dataset-dependent.

Conclusions:

• Simrank offers molecular ecologists a high-throughput, open-source solution for comparing large sequence sets.
• The utility facilitates efficient similarity searches in large biological sequence databases.