FreClu

FreClu clusters short sequencing reads de novo using a frequency-based approach to group error-derived variants into hierarchical trees to improve alignment accuracy and polymorphism detection.

Key Features:

• Tree Structure Organization: FreClu constructs hierarchical trees in which each child sequence differs from its parent by a single mutation and the root node is the most frequently observed sequence representing the reliable representative read.
• Error Correction and Alignment: By aligning the root (most abundant) sequence to the reference genome, FreClu reduces mapping of erroneous reads to false-positive positions and complements base calling and other error-correction techniques.
• Efficiency and Scalability: The algorithm operates with linear time complexity, enabling efficient processing of large short-read datasets from massively parallel sequencing technologies with deep coverage.
• Error Rate Evaluation Model: FreClu evaluates sequencing error rates using bacterial artificial chromosome (BAC) sequencing data as an internal control within the same experimental run.

Scientific Applications:

• Transcriptome and Genome Analysis: Improves read alignment accuracy to facilitate structural analyses of transcriptomes and genomes.
• Polymorphism Detection: Reduces false-positive variant calls caused by sequencing errors, improving the reliability of polymorphism detection.

Methodology:

FreClu clusters short reads by frequency into a hierarchical tree linking each sequence to a parent through potential single-base mutations; it identifies the most abundant sequence as the root and aligns that representative read to the genome; an error-rate evaluation model uses BAC sequencing data for error assessment.