Mining for Data

The progressive use of computational methods in chemistry laboratories leads to an amount of data that is barely manageable even by a team of scientists. Specialized methods in instrumental analysis, like the combination of chromatography with mass spectrometry, may produce several hundred megabytes of data in every run. Combinatorial approaches support this trend. 

These data can barely be evaluated in detail without help of computational methods in fact, it becomes more and more interesting to produce generalized information 

The task of data mining in a chemical context is to evaluate chemical data sets in search of patterns and common features to find information that is somehow inherent to the data set but not obvious. One of the differences between data mining and conventional database queries is that the characterization of chemicals is performed with the help of secondary data that are able to categorize data in a more general way and helps in finding patterns and relationships. It would be an unsuccessful approach to try to keep all potentially useful information about a chemical substance in a structure database. Thus, the extraction of relevant information from multiple data sources and the production of reliable secondary information are important for data mining. 

In the last decades methods have been developed to describe quantitative structure-activity relationships and quantitative structure-property relationships, which deal with the modeling of relationships between structural and chemical or biological properties. The similarity of two compounds concerning their biological activity is one of the central tasks in the development of pharmaceutical products. A typical application is the retrieval of structures with defined biological activity from a database. Biological activity is of special interest in the development of drugs. 

Molecular descriptors for the representation of chemical structures are one of the basic problems in chemical data mining. 

---

Text mining is the bread-and-butter method used by researchers on a daily basis [31, 32]. If you ask researchers what they really want from information management, you might be surprised how often they wish for a science Google to mine for data. As beautiful and simple as this paradigm sounds. 

---