The persistent decline of drug research and development efficiency is a familiar topic among pharmaceutical innovators. Six years have passed since Jack Scannell et al. brought attention to this fundamental challenge to the industry by pointing out that "the number of new drugs approved per billion U.S. dollars spent on R&D has halved roughly every 9 years since 1950 (falling around 80-fold in inflation-adjusted terms)." Unfortunately, today, evidence of sustainable improvements in approval rates remains elusive. Business, scientific and technical leaders across the biopharmaceutical industry continue the struggle to find new approaches that can reduce R&D costs and speed delivery of lifesaving new innovations to the patients who rely on them.
Does scientific big data hold the answer?
In recent years, business strategists have emphasized underutilization of information assets as a strong headwind impeding drug development productivity gains. The staggering number of scientific publications reinforces their hypothesis. Based on the most current available data, CAS estimates that the global body of scientific journal and patent literature comprises more than 150 million discrete publications and it's growing at a nearly exponential rate.
The challenge for researchers is finding the answers they need in this massive and diverse information haystack. In fact, a 2017 CAS survey of scientific researchers found that they spend on average 7 hours per week looking for information, with approximately 60% of that time dedicated to sorting through search results to find the relevant information. The time they invest scouring this vast domain directly impacts the time they can dedicate to developing and testing their therapeutic hypotheses. Given this, one has to wonder:
How much time and money could be saved through improved information asset utilization? What could happen to the pace of disease-target-drug hypothesis development and testing if advances in machine learning (ML) and natural language processing (NLP) can be effectively brought to bear in the pharmaceutical industry?
Amid strong industry investment, questions about teaming as a pre-requisite for success
To find answers to these questions, the (bio)pharmaceutical industry has been making significant investments in R&D information technology capital assets and expertise—including heavily recruiting data scientists—contributing to an estimated 19% growth in this occupation (including computer and information scientists) across all industries from 2016 to 2026. Leadership has also been considered, with boards of directors attempting to construct "big data" savvy executive teams. Despite these efforts and investments, relatively few success stories have been publicized to date.
Considering the focus put on each of these areas, it seems that the challenge may not be in the talent, technology or science individually, but in the ability to develop the meaningful connections between them to deliver insight in a scalable way.
Can better teaming and collaboration among discovery scientists and information technologists lead to the desired efficiency gains?
CAS, a matchmaker for sci-tech information
The challenge the biopharmaceutical industry is facing in finding better ways to harness and connect available information is not new. In fact, CAS was founded 111 years ago because chemists recognized the need for a new approach in managing the growing volume of literature in their field and making it more accessible and useful to practitioners. As the body of literature has grown and diversified over the years, and technology has advanced, CAS has evolved and developed new solutions for emerging information challenges, such as digitally storing and accessing chemical structure information, searching Markush representations within patents and rapidly analyzing large scientific data repositories. Our solutions exemplify the benefits that come from scientists and technologists working together.
The current age of big data, predictive analytics and artificial intelligence provides a new wave of opportunities and challenges to advance scientific information solutions. As the underlying information technology evolves rapidly, the unique aspects of scientific and chemical information require special attention to deliver on the promise of these technologies for biopharmaceutical use cases. CAS is uniquely positioned to help make the connection between scientific information and emerging technologies that delivers measurable impact to the R&D process and beyond.
Speeding up synthesis
Predictive synthesis is one area where CAS is uniquely bringing these technologies to bear by solving efficiency challenges in pharmaceutical R&D. Synthetic route selection and process development are complex and expensive facets of drug discovery and development. For example, medicinal and synthetic chemists sort through an estimated 100 possible choices at each synthetic juncture, en route to a target molecule. During the retrosynthetic analysis stage, they must also address a cadre of factors that impact the commercial viability of a lead compound. These factors include the realization of high reaction yields; minimization of reactor residence (cycle) times; minimization of the number of required synthetic steps; avoidance/minimization of protective group use; and the minimization of overall process energy consumption.
Predictive technology can help by analyzing a myriad of alternatives and recommending the most direct synthetic path and other possibilities based on more than 100 million reactions and related data found in the CAS content collection. This vision will be brought to life with the pending integration of ChemPlanner® in SciFindern. Together, these state-of-the-art cheminformatics technologies will help chemists find the best selection of diverse and viable routes for their syntheses—whether constructing new candidate molecules for screening or troubleshooting key transformations in preparation for scale-up into the pilot plant or commercial manufacturing setting. This is just one of many use cases for scientific big data technologies currently in development at CAS.
At CAS, we are excited to partner with research and business leaders of pharmaceutical companies to enhance the value of scientific information in the R&D process and better connect data and technology to provide insights. We look forward to finding new ways to leverage our unique content, technology and expertise to create value for our customers and help them deliver innovations that improves lives.
Have a scientific information challenge? Learn how CAS can help.