Mining and visualizing the chemical content of large databases

The shift in the drug discovery paradigm that has resulted in increased dependence on large volumes of information for decision making has created new challenges and opportunities in chemoinformatics research. The large volumes of data that are now available present an opportunity for the development of better predictive models, for the selection of more promising drug candidates, and for more comprehensive decisions to be made by researchers in chemistry and biology. The progress of this approach in drug discovery has also introduced challenges associated with the need to develop better tools to manage the large volumes of data. For chemistry-related research in particular, there is a need for more effective retrieval methods and better tools for data analysis that incorporate good interface design to facilitate the use of such systems by the non-expert scientist. This article reviews recent developments in the mining of large volumes of chemistry data, with a focus on chemoinformatics tools for small-molecule drug discovery and tools developed for analysis and in-depth mining. External sources of data, such as a variety of web services, further expand the volume of data to be analyzed during any small-molecule discovery project. The impact of the generation of large volumes of data and the related research factors is reshaping chemoinformatics work.