Identification of species based on DNA barcode using k-mer feature vector and Random forest classifier

Area of Science:

• Molecular Biology
• Bioinformatics
• Computational Biology

Background:

• DNA barcoding offers automated and accurate species identification using standardized DNA fragments.
• Developing efficient computational methods is crucial for analyzing large DNA barcode datasets.

Purpose of the Study:

• To develop a novel computational approach for species identification using DNA barcoding.
• To evaluate the performance of the proposed method against existing identification strategies.

Main Methods:

• DNA barcode sequences were transformed into numeric feature vectors based on k-mer frequencies.
• Random Forest machine learning methodology was applied for species identification.
• The approach was validated using both real and simulated biological datasets.

Main Results:

• The proposed k-mer frequency and Random Forest approach demonstrated high species identification success rates.
• It outperformed traditional similarity-based, tree-based, and diagnostic-based methods.
• Performance was comparable to existing supervised learning-based approaches.

Conclusions:

• The developed computational method provides an effective tool for DNA-based species identification.
• An online web interface, SPIDBAR, was created to make this tool accessible to taxonomists.
• This advancement aids in automated and accurate species identification in various ecological and taxonomic studies.