Mapping sequence to feature vector using numerical representation of codons targeted to amino acids for...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Phylogenetic analysis relies heavily on sequence similarity, which is computationally challenging for large datasets.
• Existing methods often require sequence alignment, which can be time-consuming and may introduce biases.
• Developing efficient and accurate alignment-free methods is crucial for advancing evolutionary studies.

Purpose of the Study:

• To propose a novel alignment-free method, CoFASA (Codon Feature based Amino acid Sequence Analyser), for nucleotide sequence similarity analysis.
• To evaluate the performance of CoFASA in phylogenetic analysis using diverse protein sequence datasets.
• To compare CoFASA's efficacy against established alignment-based and alignment-free methods.

Main Methods:

• Assigning numerical weights to nucleotides to calculate codon degrees.
• Deriving amino acid degrees based on codon degrees.
• Generating 20-dimensional sequence features using amino acid degrees and relative abundance.
• Applying these features for phylogenetic analysis of various protein families.

Main Results:

• CoFASA demonstrated comparable performance to leading alignment-based methods (e.g., ClustalW, MAFFT).
• CoFASA outperformed several well-known alignment-free methods in predicting taxonomic relationships.
• The generated features effectively distinguished sequences based on their taxonomic labels across different datasets.

Conclusions:

• CoFASA offers a cost-effective and consistent approach for phylogenetic analysis.
• The method's reliance on codon features provides a robust alternative to alignment-based techniques.
• CoFASA shows significant potential for accurate taxonomic classification and evolutionary studies.