True integrated analysis through Relational Data Mining

In drug discovery, increasingly vast amounts of data are collected on the properties of compounds, genes, cells, experimental animals, patients, etc. But current data analysis typically remains limited to a single domain: the activity of a (chemical) compound is explained in terms of its chemical properties (SAR), the behaviour of a cell in terms of properties of its (active) genes.

However, to explain why a compound is active on certain cell types but not others, explanations must take into account a combination of chemical properties of the compound and biological properties of the cells.

Moreover, traditional data mining requires that the properties used to fashion an explanation be chosen up-front and entered into a single table. This leads to loss of information (e.g. a compound cannot be reconstructed from its properties in the table) and may severely reduce the scope of possible explanations considered or even lead to erroneous explanations (if some relevant properties were left out).

PharmaDM's relational data mining technology transcends first generation methods and frees researchers from all these limitations.

Learn more about Relational Data Mining