PC-mer

PC-mer encodes DNA sequences using k-mer representations combined with nucleotide physicochemical properties to enable efficient alignment-free comparison and machine-learning classification of coronavirus genomes, including SARS-CoV-2.

Key Features:

• Encoding Methodology: PC-mer employs k-mer encoding combined with nucleotide physicochemical properties and reduces the size of encoded data by approximately 2^k times compared to classical k-mer profiling methods.
• Machine-Learning-Based Classification Tool: PC-mer applies machine learning algorithms to classify members of the coronavirus family and can process input sequences directly from the NCBI database.
• Alignment-Free Computational Comparison Tool: PC-mer computes alignment-free dissimilarity scores between coronavirus sequences at genus and species levels without sequence alignment, reducing computational demands while maintaining high accuracy.

Scientific Applications:

• Sequence Comparison: PC-mer provides dissimilarity scores for comparing sequences across coronavirus strains to support tracking viral evolution and identifying new variants.
• Phylogenetic Analysis: PC-mer's alignment-free comparisons support phylogenetic studies by offering sequence-similarity assessments as an alternative to alignment-based approaches.
• Genomic Research: PC-mer's efficient encoding and analysis methods enable handling large genomic datasets with reduced computational overhead.

Methodology:

PC-mer combines k-mer encoding with nucleotide physicochemical properties, applies machine learning classifiers for sequence classification (reported 100% accuracy), computes alignment-free dissimilarity scores and compares them to dynamic programming-based pairwise alignments (reported >98% concordance at genus level and 93% for SARS-CoV-2), and can process input sequences from the NCBI database.