Over the last two centuries, drug discovery usually begins in the chemical or pharmacological laboratory, also known as the ‘wet lab’ and is carried out by human agents, usually by the extraction of specific active compounds with hypothesised therapeutic properties from plants or animals. It could also involve the artificial synthesis of novel compounds or analogs of these active compounds.
This takes a lot of time, especially when it comes to therapy that are highly structural and require a specific three-dimensional configuration to be effective. This system is highly cost-intensive and inefficient as the researchers would have to expend time and resources in screening 1000’s of potential candidates, usually to no avail.
Finding a compound that is fitting in terms of efficacy and safety is very rare. Coupled with the cost of conducting animal research and clinical trials on the successful candidate, drug discovery using the wet lab workflow is a process that significantly contributes to not-only the length of time it takes a drug to enter the market, but also the final cost of the therapy on the consumers.
Figure 1: Timeline of Drug Discovery to Approval
The rise of digital computing in the past five decades, coupled with an exponential increase in computing power has resulted in an approach that is less ‘wet-lab’ and more computational. This approach is known as the data-driven approach to drug discovery.
It emphasises utilising very large datasets to identify drug candidates that have the highest probability of having certain therapeutic properties based on specific characteristics like molecular structure, stability, receptor-ligand interaction, catalytic properties, possible metabolic transformations in the body etc. This greatly reduces the cost of R&D at the drug development which runs into billions of dollars.
The following are the reasons we believe that data-driven drug discovery is the future:
- The Rich History of Pharmaceutical Drug Development and Research makes it a Treasure Trove: Over the past two centuries, pharmaceutical companies have generated large and unprecedented volumes of data from drug discovery and clinical research, which is rich in undiscovered and novel insights. Using big data, artificial intelligence and machine learning techniques, research companies are now able to leverage this data to push and accelerate the drug-discovery process. Astrazeneca describes in this article how the data-driven approach is being used not just in drug discovery, but also in the testing process and the validation of the results from the testing.
- Accelerated Insight Derivation: The essence of data collection in drug discovery is to derive insights that either supports or opposes the objective of the research in question. With heightened computational power, the rate at which these insights are generated is greatly increased. Pfizer already leverages technologies like Machine learning and computer vision to derive insights from their repository of data. This speaks to the rate at which data-driven drug discovery technologies are finding greater expression in the clinical research industry.
- Increased Accuracy: One undisputed advantage of computers over humans is the sustained and consistent accuracy of computer systems once the necessary parameters have been programmed. In clinical research, human error remains an unavoidable evil as humans are inherently prone to error and has led to a lot of loss. With computer systems, we can be sure of accuracy even for three dimensional biomolecular configurations which are some of the most complicated systems in existence. These computer technologies will be able to unravel and predict the potential interactions between the drug candidates and the biomolecular systems.
- Drug Repurposing: According to the National Institute for Biotechnology Interchange (NCBI), drug repurposing is the ‘Drug repurposing is using an existing drug for a new treatment that was not indicated before’. It involves different methods such as phenotypic screening, target -based techniques, knowledge-based techniques, signature-based techniques, network-based techniques and mechanism-based techniques. As described by Innoplexus, the data-driven approach can employ technologies such as ‘network analysis and machine learning, using life sciences domain ontology to help with the repurposing of drugs with minimal testing and effort’
From the above, we can see how exciting the future of drug discovery and clinical research is with the data-driven approach. Kindly share in the comments other ways the data-driven approach can advance clinical research.