Data structures based on k-mers for querying large collections of sequencing data sets

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• High-throughput sequencing generates massive datasets, often exceeding petabyte scale.
• Public repositories like the European Nucleotide Archive store these datasets for reproducibility.
• Current repositories lack efficient online search capabilities for sequence data.

Purpose of the Study:

• To provide an accessible survey of computational approaches for indexing and querying large sequencing datasets.
• To review the properties, classification, and underlying principles of these data searching methods.
• To summarize the performance, strengths, and limitations of existing techniques.

Main Methods:

• Focuses on computational approaches that represent datasets as sets of k-mers.
• Reviews and classifies various indexing and querying strategies.
• Analyzes the performance and characteristics of different methods.

Main Results:

• Several computational methods based on k-mer representation have been developed for large-scale sequence data.
• These methods offer varying performance, strengths, and limitations for indexing and searching.
• An accessible classification and overview of these approaches are presented.

Conclusions:

• Computational k-mer based methods are crucial for enabling online sequence searches in large public repositories.
• These approaches enhance data accessibility and usability for investigators.
• Further development and evaluation are needed to optimize strengths and address limitations.