Prediction of novel doping agent through the integration of chemical and biological data using in silico method

The identification of novel doping agent is not quick enough to prevent the misused of the potential doping agent among athletes. Existing techniques e.g. HPLC are costly and time consuming. Hence, other alternative methods are needed to handle this problem and one such method is in silica method. In this study, an in silica method that integrates chemical and biological data was used to predict potential doping agents. The in silica method, also known as in silica target prediction, first analyse patterns of protein-ligand binding from chemical and biological data through the use of machine learning algorithm. The deductive model built can then be used to predict potential protein targets for a novel compound, given its chemical structure. The models built were trained on compounds binding to protein targets known to produce doping effects e.g. androgen receptor which increases muscle mass obtained from ChEMBL database. This study employed three molecular descriptors (MACCS Keys, ECFP _ 4 and FCFP _ 4) and two machine learning algorithms (Decision Tree and Naive Bayes Classifier) to build the predictive model. Two validations were performed on the models which are internal and external validation. Sensitivity was used as a performance measure. The performance of the models were also analysed at a certain cut off scores and ranks. The internal validation showed that the combination of FCFP _ 4 and Decision Tree model performed best with a sensitivity value of 0.94 , when a cut off of rank 5 was applied. This model then proceeded to the external where compounds from WADA Monitoring List were subjected to testing. It was found that mitragynine and nicotine were predicted to bind to x-opioid receptor which is supported by scientific literature. This proved that the model was able to predict the protein target for compound outside of the training set. Hence, in silica method can be used to predict potential doping agents in a timely and efficient manner and consequently prevent the widespread of misused of doping agent.