Graphical representation for DNA sequences via joint diagonalization of matrix pencil

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Visual inspection of DNA sequences is crucial for understanding biological functions.
• Existing methods for sequence comparison often rely on sequence alignment, which can be computationally intensive and may miss complex relationships.

Purpose of the Study:

• To develop a novel alignment-free method for graphical representation and similarity analysis of DNA sequences.
• To provide a robust numerical characterization of DNA sequences for enhanced comparative studies.

Main Methods:

• Representing DNA sequences using a neighboring nucleotide matrix transformation.
• Employing approximate joint diagonalization theory to derive eigenvalue vectors (EVVs) as numerical sequence characterizations.
• Visualizing sequences via plotting EVVs in a 2-D plane and clustering using k-means.
• Performing similarity analyses through distance computations between EVVs.

Main Results:

• The proposed method generates unique graphical representations and numerical characterizations (EVVs) for DNA sequences.
• Clustering of sequence feature curves using k-means effectively categorizes sequences into subclasses.
• Similarity analyses based on EVV distances demonstrate the method's effectiveness, as evidenced by dendrogram construction.

Conclusions:

• The novel alignment-free approach offers a comprehensive method for DNA sequence visualization, characterization, and similarity analysis.
• This method captures more sequence information by analyzing it jointly, outperforming traditional separate analysis.
• The graphical and numerical representations facilitate a deeper understanding of sequence relationships and evolutionary patterns.