همایش ملی بیوانفورماتیک ایران

صفحه اصلی / 4th international edition and 13th Iranian Conference on Bioinformatics

Expansion and Sequencing of the DNA Code Used in the COVID-19 Vaccine Using Meta-Heuristic Algorithms

نویسندگان :

Ahmad Aliyari Boroujeni¹ Mohammadreza Parsayi² Hossein Rahmati³

1- دانشگاه زنجان 2- دانشگاه زنجان 3- دانشگاه زنجان

کلمات کلیدی :

Gene Sequence،COVID-19،Traveling Salesman Problem،Genetic Algorithm،Teaching Learning Based Optimization

چکیده :

The Spike gene of the SARS-CoV-2 virus is responsible for encoding the spike protein, which plays a key role in the virus's attachment to the ACE2 receptors on the host cell surface and its entry into the cell. This gene and its associated protein have been one of the primary targets for research in vaccine and drug development. The full length of this gene is approximately 3,822 nucleotides. However, in this part, we are using a partial sequence of 350 nucleotides. Using a shorter section (such as the first 500 base pairs) helps in quickly examining fundamental concepts, genetic patterns, or potential variations. Moreover, the initial region of the gene typically includes the start Codon (ATG) and regulatory regions that control the initiation of protein translation. This section is crucial for understanding the mechanism of protein production and for targeting in research. Using metaheuristic algorithms, we investigate the DNA sequence used in the COVID-19 vaccine by applying algorithms designed to solve the Traveling Salesman Problem (TSP) for generating, extending, sequencing, and matching other DNA sequences. The objective of the problem has been compared using four Single-crossover Genetic Algorithm (GA) (Golberg, 1989), Multi-crossover GA, Teaching-Learning Based Optimization (TLBO) (Rao et al., 2011) and New Improved TLBO (NITLBO) (Aliyari Boroujeni et al., 2023) algorithms with identical parameters, including a population size of 100 and a maximum of 300 iterations. The mutation rate for both types of GA implementations has been set to 0.7. Additionally, for the NITLBO algorithm, the weight of elitism and number of teachers are set to 0.9 and 20, respectively. In these experiments, the Single-crossover GA reached a Hamming distance of 212.5, corresponding to 60.7% accuracy. For the Multi-crossover GA, the values were 233.5 and 66.7% accuracy. The TLBO algorithm showed a Hamming distance of 331 with 94.5% accuracy, while NITLBO achieved a distance of 346 and 98.8% accuracy. The results from comparing these algorithms show that the TLBO and NITLBO algorithms perform better than the GAs. This is because in the GAs, the presence of the mutation rate parameter can cause the sequence code matching not to necessarily follow an increasing trend and although Single-crossover GA executes in a shorter time, it does not always guarantee an optimal solution.