3 Sep 2002 - DMax data mining platform: Launch version 2.1, SBS, The Hague, September 22-26

PharmaDM will participate in the 8th Annual Conference and Exhibition of the Society for Biomolecular Screening, titled "High information content screening" (September 22-26, 2002, The Hague, the Netherlands). Topics of the conference cover recent advances in screening technology, including the role of data mining. PharmaDM is co-hosting the session "From data to knowledge: Machine intelligence in informatics" (Tuesday, September 24).
PharmaDM is pleased to invite you to its technology presentation, where the company will launch the new version of its data mining platform, DMaxTM version 2.1. The newly incorporated scientific data cube provides the user with a practical way to visualise and navigate through a complex network of data.
The data cube is a common term in the context of data warehousing, where it is used to store data that can be viewed at different levels of detail or summary. Each dimension of the cube is defined by a hierarchy on top of the data that the user can use to navigate through the data in order to reach these different levels (e.g. a geographical hierarchy on top of stores, in order to summarise the sales in all stores of a certain area). PharmaDM has extended the traditional data cube functionalities drastically in order to make the data cube a practical and functional tool in life science applications:
- Both numerical data and symbolic data can be included in the cube.
- Instead of using rigid and pre-defined data hierarchies in the different dimensions, like in a traditional data cube, PharmaDM has extended the flexibility of these hierarchies significantly: they can be user-defined, or can be generated automatically, like a decision tree built in a data mining experiment.
The PharmaDM scientific data cube is an ideal means for the life science researcher to navigate through the data in an orderly way, and to view the data at all desired levels of detail. Biological, chemical and clinical information can then each represent a different dimension of the cube. A data point in the cube could for example be the expression level of a certain gene in a certain tissue (biological dimension), under the influence of a certain drug (chemical dimension) for a certain patient (clinical dimension). In this way, truly integrated data mining, combining information from different scientific domains, becomes possible.
PharmaDM is demonstrating the benefits of the data cube on a public domain data set provided within the Developmental Therapeutics Program of the National Cancer Institute (USA). This data cube contains a chemical dimension (molecules tested in the Human Tumor Cell Line Screen) and a biological dimension (cell lines described by the expression level of their genes). A well-known data mining approach in the literature is to cluster the data, hoping to find interesting molecule-target combinations. In the new PharmaDM conceptual clustering approach, biochemical rules are generated during the data mining process describing the clusters found. Scientists can thus readily extract relevant insights on the mechanism of anti-tumor action of molecules on cellular targets.

Click here for a screenshot (PDF)

Related hyperlink