skip to content

Data-Driven Drug Discovery and Molecular Informatics

My research is focused on side effects of medicines. Side effects are noxious and unintended effects of medicines [1], which can harm patients and are a significant obstacle in the development of new drugs. We use informatics and statistical methods to integrate data from different sources and discover patterns in the data. This can help us understand the cause of side effects and ultimately contribute to the development of safe medicines in the future.

Specifically, I have been studying the Food and Drug Administration Adverse Event Reporting System, which collects reports of potential side effects from doctors and patients. Since drugs interfering with human proteins can result in side effects, we use bioactivity data describing the interaction between chemical compounds and proteins from the ChEMBL database [2].

However, drug plasma concentrations, which is the concentration of drug in a patient’s blood, are also determining for the effect of drugs in the human body. As previous studies have not taken this into account on a large scale, we compiled a set of drug plasma concentrations from scientific literature and integrated these with the bioactivities. We used the ratio of the in vitro potency over the unbound drug plasma concentration to select proteins against which therapeutically-relevant plasma concentrations are reached.

We identified overrepresented combinations of adverse events and protein targets and compared these to the targets currently included on secondary pharmacology safety screens. While our results support various of these currently screened targets, our results show that in vitro bioactivity at single proteins is rarely strongly indicative of the adverse event in vivo, since bioactivities often have low precision or low recall. By quantifying the statistical associations, our work can help prioritise targets with the strongest correlation to side effects for inclusion in future drug safety screening.

 

Thanks to Lhasa Limited for funding this project.

  1. http://apps.who.int/medicinedocs/en/d/Js4893e/
  2. https://www.ebi.ac.uk/chembl/

Publications

Systematic Analysis of Protein Targets Associated with Adverse Events of Drugs from Clinical Trials and Postmarketing Reports.
IA Smit, AM Afzal, CHG Allen, F Svensson, T Hanser, A Bender
– Chemical research in toxicology
(2020)
34,
365
Network integration and modelling of dynamic drug responses at multi-omics levels
N Selevsek, F Caiment, R Nudischer, H Gmuender, I Agarkova, FL Atkinson, I Bachmann, V Baier, G Barel, C Bauer, S Boerno, N Bosc, O Clayton, H Cordes, S Deeb, S Gotta, P Guye, A Hersey, FMI Hunter, L Kunz, A Lewalle, M Lienhard, J Merken, J Minguet, B Oliveira, C Pluess, U Sarkans, Y Schrooders, J Schuchhardt, I Smit, C Thiel, B Timmermann, M Verheijen, T Wittenberger, W Wolski, A Zerck, S Heymans, L Kuepfer, A Roth, R Schlapbach, S Niederer, R Herwig, J Kleinjans
– Communications biology
(2020)
3,
573
Associating adverse drug effects with protein targets by integrating adverse event, in vitro bioactivity, and pharmacokinetic data
I Smit
(2020)
Systematic analysis of protein targets associated with adverse events of drugs from clinical trials and post-marketing reports
I Smit, A Afzal, C Allen, F Svensson, T Hanser, A Bender
(2020)
2020.06.12.135939
Information-Derived Mechanistic Hypotheses for Structural Cardiotoxicity
F Svensson, A Zoufir, S Mahmoud, AM Afzal, I Smit, KA Giblin, PJ Clements, JT Mettetal, A Pointon, JS Harvey, N Greene, RV Williams, A Bender
– Chem Res Toxicol
(2018)
31,
1119
A Comparative Analysis of Drug-Induced Hepatotoxicity in Clinically Relevant Situations
C Thiel, H Cordes, L Fabbri, HE Aschmann, V Baier, I Smit, F Atkinson, LM Blank, L Kuepfer
– PLoS Computational Biology
(2017)
13,
e1005280

Telephone number

01223 336452 (shared)