Biocomputation and Knowledge Management in Drug Discovery and Development   

Chairperson: María Isabel Loza
Universidad de Santiago de Compostela
Santiago de Compostela, Spain

Martin Stahl
F. Hoffmann-La Roche Ltd.
Basel, Switzerland
"Tools and Methods of Proven Value in Molecular Design"
A host of computational methods and tools have been developed to support the drug discovery process. This talk will highlight some of those that have been particularly useful in medicinal chemistry programs at Roche. At the same time, practical considerations of efficient use of molecular design resources will be presented. A central theme will be the use of experimental data for knowledge generation. The increasing amount and quality of available screening results, physicochemical properties, and crystal structure information represents an enormous chance to learn and to empirically improve computational prediction tools where first principles methods have stagnated. The presentation will focus on the utilization of structural databases (CSD, PDB) for improving our understanding of molecular conformations and interactions, as well as the proper use of search constraints in virtual screening.

Tudor I. Oprea
University of New Mexico School of Medicine
Alburquerque
New Mexico, USA
"From Digital to Experimental Bioactivity"
In silico technologies such as virtual screening are regarded as an increasingly reliable complement to experimental methods, in particular in the area of ligand and drug discovery. In recent years, virtual screening has demonstrated its ability to reduce the subset of the screening deck, sometimes to a number below 1000, and still converge on bioactive molecules. This talk will highlight results in the area of estrogen receptor research, in particular for GPR30, ERalpha and ERbeta, where digital experiments have translated successfully into experimental bioactivity. The workflow of the New Mexico Molecular Libraries Screening Center (http://screening.health.unm.edu), with respect to integrating virtual and biomolecular screening, will be discussed.

Angelo Carotti
University of Bari
Bari, Italy
"The Role of Structural Biology and Chemoinformatics in the Discovery of New Clinical Candidates for Neurodegeneration"
Despite the substantial investments in drug discovery research by private and public institutions over the past decades, very few drugs have reached the market and many diseases remain inadequately treated.1 Genomics, chemio- and bio-informatics, high-throughput organic synthesis and biological screening should be more efficiently integrated to overcome these objective difficulties. Following this strategy, we targeted Parkinson’s (PD) and Alzheimer’s (AD) diseases through an in silico design of two large series of safinamide and coumarin derivatives, prepared through multiple parallel synthesis, which resulted highly potent and selective MAO B inhibitors. with outstanding pharmacokinetic and toxic profiles .On the assumption that a suitable modulation of multiple targets may provide improved therapeutic effects and safer toxicological profiles, a new class of molecules acting as reversible dual MAO-B and acetylcholinesterase inhibitors was successfully designed and synthesized.

Gerhard F. Ecker
University of Vienna
Vienna, Austria
"Predicting Ligands for ABC-transporter – Structure-based vs. Ligand-based Approaches"
ABC-transporter are ATP-driven, transmembrane proteins mainly expressed in the intestine, liver, kidney, and the blood-brain barrier. They are highly polyspecific in their substrate recognition and are very often responsible for poor bioavailability and high toxicity of drug candidates. Additionally, several representatives confer multiple drug resistance in tumour therapy. Both design and inhibitors and prediction of substrate properties are therefore challenging tasks in pharmaceutical research. Within the lecture we will exemplify both structure-based and ligand-based approaches for prediction of inhibitors and substrates of selected ABC-transporter. Methods include homology modelling, unsupervised neural networks and the use of similarity based descriptors.

Manuel Pastor
GRIB, IMIM/Universitat Pompeu Fabra
Barcelona, Spain
"How Useful are the Results of Computational Chemistry Methods in Drug Discovery?"
On the 80’s, computational chemistry methods raised high expectations in the field of drug discovery and were regarded as “the way” to obtain better drugs faster. Now, twenty years and several consecutive stages of over-enthusiasm, misuse and method-bashing after we can, at last, reflect about the true potential of computational chemistry methods in focused research. Probably, one of the main reasons that hampered the practical application of many computational methods has been the way the method results were presented. Often, the endpoint of complex and valuable computation was just a graphic which was not understood by the members of the team in charge to translate these results into something useful. Not surprisingly, many of these results were ignored, the method labelled as “not useful” and their use deprecated. In this meeting we want to show some tools in development in our group that follow a alternative approach and aim to present the results of the computational chemistry methods in terms of a ranked list of synthetically accessible chemical compounds. These tools will be illustrated with the help of examples that will demonstrate that they could provide useful solutions to some practical problems in the field of drug discovery.

close window