Large chemical data sets

A most important task in the handling of molecular data is the evaluation of "hidden information in large chemical data sets. One of the differences between data mining techniques and conventional database queries is the generation of new data that are used subsequently to characterize molecular features in a more general way. Generally, it is not possible to hold all the potentially important information in a data set of chemical structures. Thus, the extraction of relevant information and the production of reliable secondary information are important topics. 

Chen X, Rusinko A, Tropsha A,Young SS. Automated pharmacophore identification for large chemical data sets. J Chem Inf Comput Sci 1999 39(5) 887-96. 

When analyzing large chemical data sets, it is common to evaluate the behavior of elements in soils, and how they change during soil formation, by dividing the mass concentration of the elements in soils by that in crustal rocks, with... 

As the wider utility of QSAR models in the environmental context is to screen, without experimental data, large chemical data sets to update priority lists, the best predictions can derive from the averaging of results from various models in a consensus approach. 

A number of studies have observed a linear correlation between precipitation or runoff and solute fluxes or chemical denudation rates (Dunne, 1978 Dethier, 1986 Stewart et ai, 2001). The relationship between solute silicon fluxes and runoff of the large watershed data set plotted in Figure 20 shows comparable relationships for granitic and basaltic rock types (Bluth and Kump, 1994). 

The exploration of large amounts of molecular data in a search for consistent patterns, correlations, and other systematic relationships can be a helpful tool to evaluate hidden information in a set of molecules. Finding adequate descriptors for the representation of chemical structures is one of the most important requirements for this task. RDF descriptors are tools for analyzing chemical data sets for molecular patterns and systematic relationships with the help of statistical analyses and... 

Analysis of large screening data sets via adaptively grown phylogenetic-like trees. Journal of Chemical Information and Computer Science, 42, 1069—1079. 

---