Ankit Kumar, Satendra Singh and Saurav Bhaskar Saha
Posters-Accepted Abstracts: Biol Syst Open Access
Background: Malaria is one of the most menacingly dreadful disease costing lives worldwide. The emergence of multiple drug resistance associated has resulted in substantial increase of its severity. Therefore, the generation of mathematical model to classify the resistant drug versus nonresistant drug is forwarding step to overcome the effect of this disease. A mathematical classifier which can classify among resistant and nonresistant drug can be an enormous aid. Methodology: The present work was carried out to exploit the previous failure experiences of resistant antimalarial drugs. Exhaustive literature search was undertaken to compile drugs with antibiotics resistance. The drugs were further manually characterized into 25 molecular and chemical attributes. The classification task was performed through 52 different classifiers using Weka machine learning work bench with 60% dataset as training and 40% as test set. Result: Out of 52 classifiers only 2 classifiers viz. SMO and VFI could reach the accuracy of 70%. Conclusion: The analysis showed that with these limited attributes and models, it is impossible to pin down these complex resistance mechanisms. Our analysis also opens a new horizon in predicting drug resistance but with different attributes and other hybrid models.
Kumar Ankit has done his B. Tech in Bioinformatics from SHIATS (Sam Higginbottom Institute of Agriculture, Technology & Sciences), Allahabad, India. His research area includes Machine learning, Mathematical modeling and he works on the development of in silico classifier for the prediction of drugs resistance in Malaria.
