Measuring Similarity among Protein Sequences Using a New Descriptor.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics and Proteomics

Background:

• Protein sequence similarity comparison is crucial for biomedical research.
• Exponential growth in protein sequence databases necessitates efficient comparison methods.
• Existing alignment-based methods are time-consuming, especially for distantly related sequences.

Purpose of the Study:

• To introduce a new, alignment-free method for protein sequence similarity assessment.
• To develop a novel graphical representation and descriptor for protein sequences.
• To evaluate the method's efficiency and accuracy across different sequence lengths and types.

Main Methods:

• Development of a unique graphical representation for protein sequences.
• Creation of a descriptor that compresses sequences into a two-value vector.
• Application and validation of the alignment-free method on beta globin, ND5, and spike protein sequences.

Main Results:

• The new alignment-free method achieves good results for both short and long protein sequences.
• The approach demonstrates significantly reduced computation time compared to traditional methods.
• Correlation and significance analyses confirm the method's effectiveness in similarity/dissimilarity assessment.

Conclusions:

• The proposed alignment-free technique provides an efficient and accurate alternative for protein sequence comparison.
• This method addresses the limitations of time-consuming alignment-based approaches.
• The graphical representation and descriptor offer a promising tool for large-scale sequence analysis in bioinformatics.