DRUG DESIGN

Data Science and Drug Repurposing
Drug Design and drug repurposing

The research and development stage of new drugs is a long, costly process with a high failure rate. Indeed, it is estimated on average that for every 10,000 molecules screened, only one will result in a Marketing Authorisation (M.A.), knowing that its development will have lasted from 10 to 15 years and will have cost, for example, in oncology, from 1.5 to 2.6 billion dollars.

One of the less time-consuming and financially costly alternatives is drug repurposing. This methodology consists of finding a new therapeutic indication for drug substances whose development has been interrupted or which have not been approved in their initial indication.

A very concrete example of the application of this method is the worldwide pandemic linked to COVID19 in recent years. Indeed, while waiting for the marketing approval of messenger RNA vaccines or other biomedical technologies, drug repurposing had its place to find already existing treatments to limit the effects of the SARSCOV2 virus in infected people.

In view of the large amount of data available in open source or private, the main difficulty is to screen a large number of drug substances in these databases and to explain the results obtained in a rational way. Data science is therefore the right tool to solve this type of problem.

Project objective

Within the framework of this project, our objective was therefore to reposition specific molecules in the fight against SARS COV2.

To do this, we had to identify pharmacological targets based on the different mechanisms of action of the coronavirus viral cycle and then select molecules with valid marketing authorisation and a known effect on the pre-identified targets.

Image reprsenting molecule and prediction
A collaborative project
This first project allowed us to work in collaboration with a multidisciplinary team comprising healthprofessionals (pharmacists, researchers, etc.) and IT professionals (developers, bioinformaticien,etc.). This scientific work was concluded with a 48-hour datathon in which the results of drug repurposing were finally presented to a variety of actors from the health world.
Pharmacological context

In order to achieve our objective, we have worked on the following elements in particular:

• Identification of interaction partners for a therapeutic target involved in a given physiological pathway.

• Identification of the active site(s) that can induce a pharmacological response.

• Screening of molecules with structural homology.

• Sorting of identified molecules according to thermodynamic and reaction variables.

• Selection of candidates with an M.A. currently validated by health organisations.

Results and conclusion

We identified 3 main molecules:

• VX-702

• Dilmapimod

• Binimetinib

Table of identified ligands
  • We confirmed our model knowing that a majority of the molecules identified had been or were being studied in clinical trials at the time of the project.
  • We sought to rationally explain the results obtained thanks to the medical and pharmacological literature and thus put forward new research hypotheses concerning certain molecules not yet studied in the fight against COVID19.

In the context of similar R&D projects within your organisation, we can deepen this type of project by crossing our theoretical results with your clinical research or medical pharmacology databases.