Promoter Prediction in DNA Classification Using Machine Learning Algorithms

The genetic code distinguishing species is stored in DNA (Deoxyribonucleic acid), consisting of four bases: Adenine (A), Cytosine (C), Guanine (G), and Thymine (T). Gene promoters, critical for controlling transcription, are located near transcription start sites. Advances in sequencing technology have accelerated DNA data growth. ML (Machine learning) has greatly aided DNA data analysis. This research aims to classify sequences as promoters or non-promoters accurately, achieving high true positive and true negative rates. We collected 106 human promoter and non-promoter data from the University of California, Irvine, ML repository., built a robust promoter classifier using various machine learning models, and utilized tokenization methods like label encoding for feature vectors with different k-mer sizes (3-mer, 6-mer). Combining 6-mer and PCA improved accuracy. Our system successfully distinguishes DNA sequences as promoters or non-promoters with high accuracy and F1-score, around 91% and 93%, automating the identification of complex biological features in sequence data.