1 Current Modeling Methods Used in QSAR / QSPR

A drug company has to ensure the quality, safety, and efficacy of a marketed drug by subjecting the drug to a variety of tests [1]. Therefore, drug development is a time-consuming and expensive process. From the initial stage of target discovery, development often takes an average of 12 years [2] and was estimated to cost USD868 million per marketed drug [3]. This high cost and lengthy process is due to the high risk of drug development failure. It was estimated that only 11% of the drugs that completed developmental stage were approved by the US or European regulators [4]. In year 2000, it was found that 10% of attrition during drug development was contributed by poor pharmacokinetic and bioavailability, while in the clinical stage, 30% of attrition was due to lack of efficacy and another 30% was caused by toxicity or clinical safety issues [4]. Thus, it will be useful to predict these failures prior to the clinical stage in order to reduce drug development costs. It was claimed that savings of USD100 million in development costs per drug could be attained with 10% prediction improvement [5]. Therefore, various methods, such as in vitro, in vivo, or in silico methods, are being used early in the drug development stage to filter out potential failures. An example of an in silicomethod is quantitative structure–activity relationship (QSAR) models, which can be used to understand drug action, design new compounds, and screen chemical libraries [6–9]. Recently, the European Chemicals Legislation, Registration, Evaluation and Authorisation of Chemicals (REACH) suggested the use of in silico methods as reliable toxicological risk assessment [10, 11]. QSARs, or quantitative structure–property relationships (QSPRs), are mathematical models that attempt to relate the structure-derived features of a compound to its biological or physicochemical activity. Similarly, quantitative structure–toxicity relationship (QSTR) or quantitative structure–pharmacokinetic relationship (QSPkR) is used when themodeling applies on toxicological or pharmacokinetic systems. QSAR (also QSPR, QSTR, and QSPkR) works on the assumption that structurally similar compounds have similar activities. Therefore, these methods have predictive and