LIFESCIENCES PULSE

From Experiments to Data-Driven Insights: The Changing Dynamics of Drug Discovery Research

 
December 21, 2018

Although a great deal of research has been done on R&D productivity, same cannot be said about research productivity limited to early stage (research to pre-clinical) of pharmaceutical R & D. However, one thing is for sure– we are on the verge of a paradigm shift in drug research. The pharmaceutical industry is slowly and steadily moving away from the old blockbuster model, and morphing into a future powered by data.

The Paradigm Shift – And Rise of Data-Intensive Research

Historically, scientific research was conducted using one of the two key methodologies: empirical which involved observation and experimentation, and theoretical which is description and explanation. The last few decades marked the beginning of the third methodology – computational science and simulations. It gained popularity as both mathematical as well as nonmathematical models could be simulated using this approach.

On the technology front, compute power got cheaper with faster supercomputers crunching massive amounts of data. Storage was also no longer a constraint as storage capacities started doubling every year. Today, the speed of genome sequencing is far exceeding the Moore’s law. The advent of high throughput technologies, data mining techniques, and omics approaches (genomics, transcriptomics, proteomics, and metabolomics) has transformed the way scientific research works.

Moving away from the norm of a single gene/protein study, the flood of data has shifted the drug discovery research paradigm and led to the emergence of the fourth methodology: data-intensive research.

Technology-Enabled Advanced Drug Research

The methods and techniques used by pharmaceuticals and biotech companies are evolving to get better drugs faster into the market. There has been a move from the ‘blockbuster one drug fits all’ approach to ‘Precision medicine’ or the right drug for the right patient. Initial steps in discovery research that used the blockbuster drug approach, required a good understanding of the disease, the targets, and structure function data on the screening compounds. Cost-effective DNA profiling and proliferation of EHR have empowered researchers to study the effect of genetic variation on the therapeutic outcome of a treatment. Pharmacogenomics, the study of how genes affect a person's response to drugs, leverages this approach to bring out drugs into the market for a select target population. This, in turn, reduces the iteration rate and paves the way for ‘precision medicine.’

Both the blockbuster drug approach and the pharmacogenomic approach are developed using target-oriented drug design. Biological complexity that is responsible for the differences within individuals is not considered. Efficacy testing is conducted for the target under consideration. FDA approved drugs, however, are known to bind to multiple targets in the body at varying concentrations. Shutting down one pathway, therefore, may result in the activation of alternate pathways in the complex biological system. This may act as a significant factor contributing to drug recall or failures detected at the clinical trial stage. However, a better understanding of biological processes can change this if new-age drug research is guided by information on pathways and phenotypes in addition to information on target and drug.

Unlocking the True Potential of Data

Tapping the full potential of data within an organization as well as across external sources can uncover several hidden possibilities. Omics data available across multiple scientific repositories, clinical data, published research data, and real-life data can deliver valuable insights and establish relationships that can enhance and accelerate the drug research process. For example:

· Drug screening can be carried out on multiple potential targets using phenotypic drug discovery strategies rather than on a single target.

· Identifying cross binders earlier on in the drug discovery process is now possible with novel ways of studying drug-target interactions using a network approach and deep learning.
 
· Al and NLP can be used for the identification of novel targets, repositioning of existing drugs, treatment of cancers, and other disorders.

 

Collaboration is the Key

Unfortunately, pharmaceutical companies today are not equipped to make this change as there is a paucity of in-house skill sets to adopt new age technologies. One way to meet the changing needs is –collaboration.

While collaborating internally extends knowledge base and leads to insights across portfolios, external partners can scale capabilities and provide access to expertise and data. Academic collaboration, on the other hand, can be a source of innovation and enables insight sharing.

It’s time for companies to move away from data silos, adopt the data-driven approach, and leverage technologies that promote collaboration such as cloud. Embracing cloud enables collaborations and data accessibility without compromising on security and provides a cost-effective on-demand compute and storage.

The pharmaceutical industry is continually progressing, and there’s no doubt that these trends will keep evolving as well. However, as we morph into the year 2019, one thing is for sure – AI and ML enabled data-driven approach is ready to usher in a whole new era of drug discovery research.

 

Arundhati Saraph works in the area of discovery research and has over 23+ years of experience in Biomedical and Life Sciences research.  Arundhati has a PhD in Biochemistry from National Chemical Laboratory (NCL), Pune. She worked as a Postdoctoral Fellow at Rensselaer Polytechnic Institute, (RPI) Troy, NY, USA, where she looked after the design and validation of peptide-based therapeutics.  She has experience in leading research projects, collaborations, and partnerships with industry and research organizations that have resulted in several publications and development of solutions that can be leveraged for bio-medical research and the pharma sector.