Exploiting fuzzy weights in CNN model-based taxonomic classification of 500-bp sequence bacterial dataset

Area of Science:

• Microbiology
• Bioinformatics
• Computational Biology

Background:

• Accurate bacterial taxonomic classification is vital for understanding microbial diversity and evolution.
• Classifying short bacterial DNA sequences (500-bp) presents significant challenges for traditional methods.

Purpose of the Study:

• To develop an improved Fuzzy-weighted Convolutional Neural Network (F-CNN) for precise bacterial DNA sequence classification.
• To address the limitations of existing methods in classifying short DNA segments and samples with similar probabilities.

Main Methods:

• Implementation of a deep learning model integrating fuzzy logic for enhanced classification.
• Inclusion of a feature selection stage and a fuzzy weighting system to manage classification uncertainty.
• Utilizing the Ribosomal Database Project Release 11 (RDP 11) dataset for model training and evaluation.

Main Results:

• The proposed F-CNN model achieved a classification accuracy of up to 84.03% at the genus level for 500-bp bacterial DNA segments.
• Demonstrated superior performance compared to traditional methods, especially for short sequence regions.
• Exhibited high generalization capabilities when applied to longer DNA sequences.

Conclusions:

• The improved F-CNN offers a robust solution for bacterial taxonomic classification using limited DNA sequence data.
• This advancement has significant implications for microbiology, epidemiology, and environmental science research.
• Accurate bacterial classification is essential for ecological and disease outbreak investigations.