iRSpot-Pse6NC2.0

iRSpot-Pse6NC2.0 predicts recombination hotspots in Saccharomyces cerevisiae genomes using hexamer features encoded by Pseudo K-tuple Nucleotide Composition combined with Support Vector Machine classification to support analysis of meiotic double-strand break initiation and genome evolution.

Key Features:

• Predictive model: Uses a Support Vector Machine (SVM) classifier that integrates hexamer features into Pseudo K-tuple Nucleotide Composition (PseKNC).
• Feature selection: Applies a binomial distribution feature selection approach to identify informative hexamer features.
• Dataset composition: Trains on datasets containing both open reading frame (ORF) and non-ORF recombination sites.
• Validation: Validated by 5-fold cross-validation with a reported maximum overall accuracy of 77.61%.
• Comparative assessment: Evaluated against 15 existing computational methods by comparing algorithms, extracted features, and predictive capabilities.
• Benchmarking: Benchmarked on chromosome XVI of S. cerevisiae and tested on independent datasets to assess generalization.

Scientific Applications:

• Recombination hotspot mapping: Identification of hotspot locations to investigate meiotic double-strand break (DSB) initiation and recombination landscape in S. cerevisiae.
• Genome evolution and diversity: Facilitates studies of genetic diversity and the role of recombination in genome evolution.
• Method benchmarking: Provides a reference for comparing computational approaches to hotspot identification.

Methodology:

Hexamer features were encoded using PseKNC and selected via a binomial distribution approach, then classified with a Support Vector Machine (SVM); performance was assessed by 5-fold cross-validation, comparison with 15 existing methods, benchmarking on chromosome XVI, and independent dataset testing.