Abstract
One of the most intensive areas of research within the pharmaceuticalindustry today is to collect and analyse data on absorption,distribution, metabolism, excretion and toxicity (ADMET). The main purpose of this is to guide drug development projects in the search for promising compounds as well as to detect, as early as possible, compounds that should not be put forward as drug candidates. Expensive synthesis and test procedures can often be avoided by instead using computer simulations, something which often is referred to as in silico modelling. Information fusion has become a well-established term for the research field concerned with methods for automatically or semi-automatically transforming information from multiple sources into a form that provides effective support for human or automated decision-making. The overall focus of this project is on information fusion for in silico modelling by combining information from multiple sources of molecular information, such as different sets of molecular descriptors, experimental data on various ADMET properties as well as libraries of molecules that are manually annotated in free text format. The aim is to enable generation of more reliable in silico models, as well as enabling chemists to understand the reasons behind predictions made by the models, thus ultimately allowing discovery of new factors for the ADMET properties. Besides addressing the industrial and scientific research problems, one main objective of the project is to establish long-term research collaboration between University of Skövde, University College of Borås, AstraZeneca and Lexware Labs. From the academic side, the researchers are expected to benefit from getting access to industrially relevant problems and non-synthetic data. The companies are expected to benefit from direct access to researchers developing and evaluating new technology that potentially have a significant impact on the industrial processes.