Tracking Evolutionary Links Among Coronavirus Types Using Self-Organizing Neural Networks

The Coronaviruses are worldwide in distribution and caused an epidemic in China in 2003. The rapidly mutating Coronaviruses spreads fast by taking different forms and infects not only human beings but also cattle, pigs, rodents, cats, dogs and birds. In fighting against these viruses, it is important to elucidate the evolutionary links among the different types of Coronaviruses. This article explores the possible roots of the evolution of Coronaviruses with the help of unsupervised Self-Organizing Map (SOM) neural network. The migration paths of these viruses are analyzed with different Self-Organizing maps based on the different genomic signatures of 50 complete Coronavirus genomes. The results are corroborated with other findings and thus Self-Organizing maps are proved to be fast, efficient, and economical tool to use as initial pointers to detailed phylogenetic analysis. New subgroups are also revealed by clustering these genomes with SOM.