DNAshapeR

DNAshapeR predicts DNA shape features from nucleotide sequences or genomic coordinates to enable quantitative analysis of DNA structural attributes for genomic and machine learning studies.

Key Features:

• High-throughput DNA shape prediction: Performs ultra-fast, high-throughput prediction of multiple DNA shape features from input sequences or genomic coordinates.
• Input formats: Accepts nucleotide sequences and genomic coordinates derived from genomic sequencing data as inputs.
• Graphical outputs: Produces graphical representations of DNA shape features for visualization and exploratory analysis.
• Feature encoding into k-mer/shape matrices: Encodes DNA sequence and shape features into user-defined combinations of k-mers and DNA shape attributes to generate feature matrices compatible with machine learning software.
• Implemented in R: Implemented in the statistical programming language R.

Scientific Applications:

• Genomic structural analysis: Supports genomic studies that require prediction and visualization of DNA structural features.
• Machine learning feature generation: Provides feature matrices combining sequence and shape attributes as inputs for machine learning models in genomics.
• Statistical learning and modeling: Supplies quantitative DNA shape attributes for statistical learning applications to study effects of DNA structure on biological processes.

Methodology:

Takes nucleotide sequences or genomic coordinates as input, applies algorithms to predict DNA shape features, and encodes predictions into matrices that combine sequence k-mers with DNA shape attributes.