Machine Learning Prediction of 3D Domain Swapping Proteins in Medicinal Plants

3D domain swapping is a protein structural phenomenon in which two or more protein subunits exchange identical structural subunits and form oligomers. Proteins that exhibit 3D domain swapping play a crucial role in various biological functions, such as secondary metabolite biosynthesis, and in coping with several biotic and abiotic stresses in medicinal plants. This study investigates the ability to predict 3D domain swapping patterns among the genomes of medicinal plants using random forest and K-nearest neighbor classifiers models, demonstrating accuracies of 91.6% and 88.7%, respectively. A total of 420 (31%) of sequences were predicted as being putatively involved in 3D domain swapping. An enrichment investigation was also carried out on the predicted 3D domain-swapped protein sequences from various medicinal plants for function annotation based on Gene Ontology (GO) terms, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways analysis, and their domain distribution in secondary metabolites biosynthesis pathways. Functional annotation of predicted sequences infers that 3D domain swapped sequences were involved in diverse molecular functions such as photosynthetic electron transport in photo system II and electron transporters, transferring electrons within the cyclic electron transport pathway of photosynthesis activity, oxidative phosphorylation, and gene regulation of environmental stresses (biotic and abiotic) by synthesizing secondary metabolites (terpenoids, alkaloids, and polyamines). These findings underscore the ability of machine learning to predict the involvement of proteins in the 3D domain-swapping phenomenon, their respective function, and their potential to facilitate drug discovery and bioengineering initiatives.