FSBC: fast string-based clustering for HT-SELEX data.

Area of Science:

• Bioinformatics
• Molecular Biology
• Computational Chemistry

Background:

• High-throughput systematic evolution of ligands by exponential enrichment (HT-SELEX) combined with deep sequencing identifies aptamer candidates from vast oligonucleotide libraries.
• Current clustering methods for HT-SELEX data often rely on full-length sequences or fixed-length motifs, limiting the identification of aptamer binding regions of varying lengths.
• There is a need for clustering methods that can analyze variable-length target binding regions and efficiently process large datasets from a single round of HT-SELEX.

Purpose of the Study:

• To develop a novel clustering method for HT-SELEX data that considers variable lengths of target binding regions.
• To create a computationally efficient clustering method that can process large datasets from a single round of HT-SELEX.
• To improve the accuracy of identifying aptamer candidate clusters.

Main Methods:

• Developed Fast String-Based Clustering (FSBC), a method designed to identify clusters by searching for over-represented strings of various lengths, representing potential target binding regions.
• FSBC incorporates search space reduction and is optimized for fast computation using data from a single round of HT-SELEX, accounting for imbalanced nucleobases.
• Compared FSBC's calculation time and clustering accuracy against four conventional methods (FASTAptamer, AptaCluster, APTANI, AptaTRACE) using over 15 million HT-SELEX sequences.

Main Results:

• FSBC demonstrated high clustering accuracy, outperforming other methods in identifying relevant sequence groups.
• FSBC achieved the second-fastest calculation speed among the compared methods, indicating significant computational efficiency.
• FSBC, AptaCluster, and AptaTRACE successfully completed clustering for the entire dataset, with FSBC and AptaTRACE showing superior accuracy.

Conclusions:

• FSBC is a highly accurate and efficient clustering method applicable to large HT-SELEX datasets.
• The method facilitates the precise identification of aptamer candidate groups by considering variable-length binding regions.
• FSBC can significantly aid in the aptamer discovery pipeline by improving the analysis of HT-SELEX data.