Drug discovery informatics
The Team develops and supports EMBL-EBI’s database - ChEMBL - www.ebi.ac.uk/chembl - containing quantitative small molecule bioactivity drug discovery data.
Although great progress has been made in developing biological drugs, synthetic small molecule and natural product-derived drugs still form the majority of novel life-saving drugs. The process complexity and costs of discovering new drugs has recently risen to the point where conventional industrial R&D efforts are being complemented with public-private partnerships. Central to this public sector activity is data sharing and availability of structure, binding, functional and ADMET data for large sets of compounds). The ChEMBL database stores curated 2-D chemical structures and abstracted quantitative bioactivity data alongside calculated molecular properties. The majority of the ChEMBL data is derived by manual abstraction and curation from the primary scientific literature, and therefore cover a significant fraction of the structure–activity relationship (SAR) data for the discovery of modern drugs. We complement this data with depositions, and selected data from other public resources, such as PubChem. Our associated research interests focus on data-mining ChEMBL data applied to drug-discovery challenges.
The underlying theme of our research interests is to perform data-mining on the ChEMBL database, applied to translational drug discovery.
Current Projects
Differential Drug Response as a Function of Age and Gender
Medicines are generally developed to be effective and safe in adult populations, and only a small subset of drugs has been formally studied for efficacy and safety in paediatric and geriatric populations. However, it is not easy to extrapolate doses, therapeutic indications, safety and efficacy, from one age group to another. Physical size, weight, fat fraction, metabolic state, drug target expression, etc., are all factors that play a role in drug absorption, distribution, and ultimately efficacy and elimination, and they change across the life span. However, the lack of understanding of how these factors change, especially at a molecular level, has been a hurdle in the transfer of medicines from one population to another, giving rise to a few documented cases of differential drug responses: increase number of side effects or lack of efficacy. The aim of this project is then, to try and identify molecular differences between differently aged human populations, which could increase the understanding of some of the age-related drug differential responses reported in the literature, and also to identify how these changes translate to species commonly used in drug development, in order to establish the best age-driven animal models.
Chemogenomic Characterization of Allosteric Modulators
Allosteric modulators are interesting compounds from the perspective of a medicinal chemist. Their relevance for drug discovery comes mostly from their mode of action, which keeps intact normal physiological regulation mechanisms, and the fact that they allow modulation of proteins not easily accessible to orthosteric modulation, e.g. peptide binding proteins. Moreover, it has been shown that they are specific ligands that possess a self limiting on-target activity. Yet, these qualities make allosteric modulators a class of compounds that is very diverse with an equally diverse array of modes of action. While several allosteric drugs are on the market and a large number of allosteric compounds are being studied in preclinical research, there is no definitive course of action to discover a novel allosteric modulator of a pathologically relevant drug target.
This project tries to capture what makes compounds allosteric modulators, but also tries to connect these observations to the targets the compounds modulate. Using the ChEMBL database and included target annotation the work leads to predictive models that can quantitatively predict the likelihood of compounds to modulate a certain target. Hence this opens the door for elucidation of mode of action, focussed libraries, and screening prioritization.
The Functional Therapeutic Chemical Classification System (FTC)
The mode of action of a compound (example: 'Anti-blood coagulation agent') is a core notion in drug discovery. It's the starting point to characterise what the compound is doing in a body. The use or therapeutic indication of the drug will logically depend on its mode of action. This abstract concept needs however to be defined and characterised in order to be used for large-scale computational analysis for instance. In order to fulfill this need, we are currently developing the Functional Therapeutic Chemical Classification System (or FTC in short). The FTC is a resource formalising the mode and mechanism of action of drugs using semantic technologies. This classification has been designed to support drug re-purposing and can be seen as a toolbox designed to fix dysfunctioning biological systems.
Future plans
We will continue our focus on translational and safety biology, with extension to areas including the study of the impact of human genetic and physiological variation of drug efficacy and safety.
Recruitment
All our research topics offer the opportunity to study and develop cutting-edge Open Source drug discovery technologies in a highly collaborative, diverse and international setting. If we have any positions they will be listed on the EMBL recruitment page, although we also welcome contact with people interested in working with us in the future
Internships
We regularly host interns within the group, you should have good programming skills, and experience in either bioinformatics or chemoinformatics. If you are interested, please send a current, full cv to kholmes (at) ebi.ac.uk