Protein active site prediction for early drug discovery and designing

Abstract Adenosine triphosphate (ATP) is an energy compound present in living organisms and required by cells for performing operations such as replication, molecules transportation, chemical synthesis, etc. ATP connects with through specialized sites called ATP-sites. ATP-sites are various proteins of a cell. The life span cell can be controlled controlling compounds without the provision to compounds, cannot survive. Countless diseases treatment (such cancer, diabetes) possible once protein active predicted. Considering need algorithm that predicts higher accuracy effectiveness, this research work very novel way. Till now Position-specific scoring matrix (PSSM) along many physicochemical properties have been used features deep neural networks order create model To overcome problem complex computation, exertion proposes k-mer feature vectors simple machine learning (ML) models attain same or even better performance less computation required. Using 2-mer vectors, trained tested five different including KNN, Conv1D, XGBoost, SVM Random Forest. gave best on features. created 96%, MCC 90% ROC-AUC 99%, which some aspects than state-of-the-art results. results 97%, 78% 92%. One benefits it much simpler more accurate.