An integrated data management framework for drug discovery--from data capturing to decision support.

Drug discovery is a highly complex process requiring scientists from wide-ranging disciplines to work together in a well-coordinated and streamlined fashion. While the process can be compartmentalized into well-defined functional domains, the success of the entire enterprise rests on the ability to exchange data conveniently between these domains, and integrate it in meaningful ways to support the design, execution and interpretation of experiments aimed at optimizing the efficacy and safety of new drugs. This, in turn, requires information management systems that can support many different types of scientific technologies generating data of imposing complexity, diversity and volume. Here, we describe the key components of our Advanced Biological and Chemical Discovery (ABCD), a software platform designed at Johnson & Johnson to bring coherence in the way discovery data is collected, annotated, organized, integrated, mined and visualized. Unlike the Gordian knot of one-off solutions built to serve a single purpose for a single set of users that one typically encounters in the pharmaceutical industry, we sought to develop a framework that could be extended and leveraged across different application domains, and offer a consistent user experience marked by superior performance and usability. In this work, several major components of ABCD are highlighted, ranging from operational subsystems for managing reagents, reactions, compounds, and assays, to advanced data mining and visualization tools for SAR analysis and interpretation. All these capabilities are delivered through a common application front-end called Third Dimension Explorer (3DX), a modular, multifunctional and extensible platform designed to be the "Swiss-army knife" of the discovery scientist.

Using particle swarms for the development of QSAR models based on K-nearest neighbor and kernel regression.

We describe the application of particle swarms for the development of quantitative structure-activity relationship (QSAR) models based on k-nearest neighbor and kernel regression. Particle swarms is a population-based stochastic search method based on the principles of social interaction. Each individual explores the feature space guided by its previous success and that of its neighbors. Success is measured using leave-one-out (LOO) cross validation on the resulting model as determined by k-nearest neighbor kernel regression. The technique is shown to compare favorably to simulated annealing using three classical data sets from the QSAR literature.

On the use of neural network ensembles in QSAR and QSPR.

Despite their growing popularity among neural network practitioners, ensemble methods have not been widely adopted in structure-activity and structure-property correlation. Neural networks are inherently unstable, in that small changes in the training set and/or training parameters can lead to large changes in their generalization performance. Recent research has shown that by capitalizing on the diversity of the individual models, ensemble techniques can minimize uncertainty and produce more stable and accurate predictors. In this work, we present a critical assessment of the most common ensemble technique known as bootstrap aggregation, or bagging, as applied to QSAR and QSPR. Although aggregation does offer definitive advantages, we demonstrate that bagging may not be the best possible choice and that simpler techniques such as retraining with the full sample can often produce superior results. These findings are rationalized using Krogh and Vedelsby's decomposition of the generalization error into a term that measures the average generalization performance of the individual networks and a term that measures the diversity among them. For networks that are designed to resist over-fitting, the benefits of aggregation are clear but not overwhelming.

Feature selection for structure-activity correlation using binary particle swarms.

We present a new feature selection algorithm for structure-activity and structure-property correlation based on particle swarms. Particle swarms explore the search space through a population of individuals that adapt by returning stochastically toward previously successful regions, influenced by the success of their neighbors. This method, which was originally intended for searching multidimensional continuous spaces, is adapted to the problem of feature selection by viewing the location vectors of the particles as probabilities and employing roulette wheel selection to construct candidate subsets. The algorithm is applied in the construction of parsimonious quantitative structure-activity relationship (QSAR) models based on feed-forward neural networks and is tested on three classical data sets from the QSAR literature. It is shown that the method compares favorably with simulated annealing and is able to identify a better and more diverse set of solutions given the same amount of simulation time.