Applying our technology on chemogenomics data will allow you to tackle some common problematics of the pharmaceutical industry:
- Structural alert detection and toxic endpoints determination.
- Hit identification, lead optimization and side effects anticipation.
- Selective or polypharmacological profile extraction.
To figure out how our technology works on chemicals data, we have selected the following mutagenicity example.
The input is a database of chemicals partitioned into two subsets: one with mutagenic chemicals and another one with non-mutagenic ones. The aim of the study is to extract the (combinations of) fragments affecting the mutagenicity of the chemicals. Consequently, the first step is to describe the chemicals by using their physical structures. Despite the complexity and the variety of chemical structures, our technology provides users with the freedom to choose the descriptors used to describe the chemicals: from atoms pairs to combinations of high level fragments.
Once all the chemicals of the database have been described with the same set of descriptors, our algorithm highlights the most contrasting (combinations of) fragments by using data mining or statistical constraints such as frequency, emergence or statistical significance.
The first way to analyse the output results is to take each (combination of) fragment(s) alone and to discuss its contrast into the database and its influence on mutagenicity. However, our output results are fully understandable and more information can be deduced by cross-referencing the (combinations of) fragments together.
QUIID provides a way to navigate into the set of (combinations of) fragments
. To have more information about the use of KAD on drug discovery problematics, please visit the kad web site
or the demos page