Classification of Protein Sequences by a Novel Alignment-Free Method on Bacterial and Virus Families

Area of Science:

• Bioinformatics
• Computational Biology
• Sequence Analysis

Background:

• Protein sequence classification is crucial in bioinformatics.
• Current sequence alignment methods are time-consuming for large datasets.
• There is a need for efficient protein classification techniques.

Purpose of the Study:

• To propose a novel method for protein sequence classification.
• To reduce the time cost of protein sequence clustering without compromising accuracy.
• To offer an alternative to traditional sequence alignment algorithms.

Main Methods:

• Developed an accumulated natural vector method.
• Projected protein sequences into a 250-dimensional space based on amino acid distribution.
• Utilized Euclidean distance to measure biological distances between protein sequences.
• Applied convex hull analysis for classification and verification.

Main Results:

• The accumulated natural vector method achieved protein sequence clustering at a lower time cost.
• Accuracy was maintained compared to existing methods.
• The method demonstrated robust performance on virus and bacteria datasets.

Conclusions:

• The proposed accumulated natural vector method is an effective and efficient approach for protein sequence classification.
• This method offers a valuable tool for bioinformatics research, particularly for large-scale sequence analysis.
• The technique provides a scalable solution for understanding protein relationships and functions.