Using ChemAxon tools in Accamba, a project for modelling screening results based on machine learning methods

poster · 15 years ago
by Dragos Horvath, Samia Aci, Céline Charavay, Samuel Wieczorek, Eric Maréchal, Gilles Bisson, Sylvaine Roy, Laurence Lafanechère (Commissariat à l’énergie atomique (CEA))
Screen Standardizer JKlustor JChem Base

The Accamba project ( is a collaborative multidisciplinary project supported by the French ministry of research, involving biologists, chemists and computer scientists. Its main objective is to develop tools in order to analyse chemical libraries and to model screening results using Machine Learning approaches. In this poster, we present a general overview of the Accamba approach and we focus on the benefits brought by using ChemAxon software tools in various stages of this project. We use Marvin beans in our screening results analysis software, Standardizer to detect automatically chemical errors and redundancies when we receive molecules databases, to clean them and normalize molecules, Calculator Plugins to compute relevant molecular descriptors, API from Screen and Klustor to make comparative studies with the new distance algorithm that we are developing in this project.In the future, we would like to allow the biologist user to better exploit our screening results database by adding structural search using ChemAxon Tools (JChem Base, Marvin beans…). We also think to use Reactor to investigate about the advantages of a virtual synthesis stage to help chemists in the Accamba project.