Not only does Intelligent Pharma offer several chemoinformatics and bioinformatics services to its costumers, but we also develop high-level research projects in new computational technologies and determine their application in meeting the customers' biomedical requirements. Intelligent Pharma is developing the following research projects with grants from CIDEM and Fundació Catalana per a la Recerca (The Catalan Foundation for Research and Innovation).
Projects
Flexible QSAR: Classical QSAR has typically shown high error rates in actual drug design projects. At Intelligent Pharma we have identified some of the reasons why such failures occur and we propose several mathematical innovations to deal with this issue. The main mathematical and computational topic that Intelligent Pharma scientists are researching is the development of functional molecular descriptors and customized machine learning tools that are able to deal with functional data.
In classical QSAR projects, molecular descriptors are fixed values for each molecule of a training set, i.e. molecular weight, isoelectrical point, etc. New functional molecular descriptors being developed by Intelligent Pharma are not constant values, but functions that depend on the internal flexibility of the molecules, for instance, a molecular descriptor can be a function that describes the volume of every molecule depending on its degrees of internal flexibility.Docking: Intelligent Pharma is committed to substantially improving high-throughput docking performance. Not only are we doing research into ligand-protein thermodynamics, but also we are trying to design an optimal fully-automated protocol.
The whole docking procedure is divided into three main points: the preparation of both ligand and receptor; the docking itself; and the analysis of results. All of them must be optimized in order to increase performance of the entire fully-automated process. At present, highly-reliable results can be reached for single cases in which much human intervention is required. However, a higher quality process is still required when a great number of cases are analyzed, particularly when they include different receptors, such as in virtual target profiling.Three-dimensional Molecular Alignment: Three-dimensional alignment of active compounds is a good tool to derive three-dimensional pharmacophore patterns. For such a reason, Intelligent Pharma is actively researching new alignment methods, based on evolutionary computation. The two issues that we are addressing are the development of parallel evolutionary algorithms and new evaluation methods that determine the quality of an alignment.
The evaluation methods that we are developing are not based on structural molecular features, but on the three-dimensional distribution of physico-chemical molecular properties.Artificial Intelligence and Hit Identification from Medicinal Plants: Products found in nature play a major role as starting material for drug discovery. For instance, of the 974 new small-molecule chemical entities found in approved drugs from 1981 to 2006, 63% were naturally derived or nature-inspired. However, despite the enormous potential of natural products, only a minor fraction of the globe’s living species have ever been tested for any bioactivity. For this reason, Intelligent Pharma has a research project that involves hit identification of natural products, ethnobotanics, and artificial intelligence.
|
|
|