An Overview of Data Mining Algorithms in Drug Induced Toxicity Prediction
Ankur Omer, Poonam Singh, N.K. Yadav and R.K. Singh
Affiliation: Division of Toxicology, CSIR-Central Drug Research Institute, Lucknow, India.
Keywords: Bioinformatics, computational prediction, data mining, in silico, machine learning, toxicity prediction.
The growth in chemical diversity has increased the need to adjudicate the toxicity of different chemical
compounds raising the burden on the demand of animal testing. The toxicity evaluation requires time consuming and
expensive undertaking, leading to the deprivation of the methods employed for screening chemicals pointing towards the
need to develop more efficient toxicity assessment systems. Computational approaches have reduced the time as well as
the cost for evaluating the toxicity and kinetic behavior of any chemical. The accessibility of a large amount of data and
the intense need of turning this data into useful information have attracted the attention towards data mining. Machine
Learning, one of the powerful data mining techniques has evolved as the most effective and potent tool for exploring new
insights on combinatorial relationships among various experimental data generated. The article accounts on some
sophisticated machine learning algorithms like Artificial Neural Networks (ANN), Support Vector Machine (SVM),
k-mean clustering and Self Organizing Maps (SOM) with some of the available tools used for classification, sorting and
toxicological evaluation of data, clarifying, how data mining and machine learning interact cooperatively to facilitate
knowledge discovery. Addressing the association of some commonly used expert systems, we briefly outline some real
world applications to consider the crucial role of data set partitioning.
Rights & PermissionsPrintExport