Importance of Applicability Domain of QSAR Models

Importance of Applicability Domain of QSAR Models

Kunal Roy (Jadavpur University, India) and Supratik Kar (Jadavpur University, India)
Copyright: © 2017 |Pages: 32
DOI: 10.4018/978-1-5225-1762-7.ch039
OnDemand PDF Download:
List Price: $37.50


Quantitative Structure-Activity Relationship (QSAR) models have manifold applications in drug discovery, environmental fate modeling, risk assessment, and property prediction of chemicals and pharmaceuticals. One of the principles recommended by the Organization of Economic Co-operation and Development (OECD) for model validation requires defining the Applicability Domain (AD) for QSAR models, which allows one to estimate the uncertainty in the prediction of a compound based on how similar it is to the training compounds, which are used in the model development. The AD is a significant tool to build a reliable QSAR model, which is generally limited in use to query chemicals structurally similar to the training compounds. Thus, characterization of interpolation space is significant in defining the AD. An attempt is made in this chapter to address the important concepts and methodology of the AD as well as criteria for estimating AD through training set interpolation in the descriptor space.
Chapter Preview


Quantitative structure-activity relationship (QSAR) modelling has become an important tool in the new drug candidate design, environmental fate modeling, toxicity and property prediction of chemicals and pharmaceuticals since they offer an economical and time-effective alternative to the medium throughput in vitro and low throughput in vivo assays (Perkins et al., 2003; Selassie, 2003; Walker et al., 2003). A QSAR model is a simple mathematical equation that is evaluated from a set of molecules with known activities/properties/toxicities using computational approaches. Thereafter, the developed predictive QSAR models are also applied by regulatory agencies to evaluate physical, chemical, and biological properties of individual chemical entities using applications specific for decision-making frameworks in risk and safety assessments (Kar & Roy, 2010). The QSAR modeling hypothesis also supports the 3Rs (replacement, refinement and reduction in animals in research) paradigm due to an increased pressure from social and economic background to trim down the use of animal testing as an important alternative method for future prediction of untested chemical entities (Benigni & Giuliani, 2003).

Robust validation of QSAR models plays a key step for the selection of a predictive model that may be considered for future prediction of new molecules. As numerous numbers of researches have been directed to the design of new molecules with the utilization of QSAR technique, validation of a QSAR model has been certified as the most considerable stride (Carlsen et al., 2009) for assessing the quality of data, applicability and mechanistic interpretability of the developed model. Thus, a huge number of investigations are currently directed towards the introduction of more appropriate validation approaches for more accurate and predictive QSAR model development. One important objective of QSAR modeling is to predict activity/property/toxicity of new chemical entities falling within the applicability domain of the developed models. The reliability of any QSAR model depends on the confident predictions of these new compounds based on applicability domain (AD) of the modeled compounds and therein arises the importance of the AD study (Ferreira, 2001).

In order to establish the scientific validity of a QSAR model and to facilitate its acceptance for regulatory purposes, the Organization for Economic Cooperation and Development (OECD) in its joint meeting (OECD, 2007) has agreed to five principles that should be followed during the construction of QSAR models. The OECD Principle 3 which defines the need of an AD expresses the fact that QSARs are unavoidably associated with restrictions in terms of the types of chemical structures, physicochemical properties and mechanisms of action for which the models can generate reliable predictions. The applicability domain of a QSAR model has been defined as the response and chemical structure space, characterized by the properties of the molecules in the training set. The developed QSAR model can predict a new compound truly only when it falls within the applicability domain of the developed model (Netzeva et al., 2005). Thus, to identify the interpolation (true prediction) or extrapolation (less reliable prediction) of query compounds is an important task for a QSAR model developer using the information of applicability domain (Netzeva et al., 2005).

Viewing the importance of AD in QSAR model validation, we herein focus to get an overview of different traditional as well as relatively new AD approaches used to judge the quality of the QSAR models. This book chapter will be helpful for the QSAR learners in order to have a clear idea on the principles of available AD approaches useful for judging the predictive quality of QSAR models.

Complete Chapter List

Search this Book: