Posted on April 17th, 2019 by Frederik van den Broek in Pharma R&D
Machine learning and deep learning, which fall under the umbrella of AI, are hot topics in pharma R&D today—and they are here to stay. “Every day, chemists are working on the design, synthesis, and application of various molecules and formulations,” explains Elsevier’s Matt Clark in R&D. “Finding hidden relationships for classification and discovery work is exactly what deep learning AI excels at, thereby freeing up researchers’ precious time. Additionally, since machines can work far faster than humans around the clock, deep learning could well find alternative and more efficient synthesis routes to create new molecules, allowing for even greater savings.”
Though some of the hubbub is still more about the tremendous possibilities than the actual realities of machine learning, there has been ongoing progress, particularly due to advances in data analytics and technological breakthroughs. For instance, facial recognition technology is now being used to help scientists rapidly analyze images of cells.
In a piece in Healthcare Weekly, Iolanda Bulgaru highlights global pharma giant Novartis’ use of deep learning to mimic the way our eyes and brains process photographic information. “The human eyes use highly connected neural networks to transform perceived light patterns into shapes and colors we associate with familiar faces, objects, and other things around us,” she writes. “Borrowing a leaf from this, the folks at Novartis taught a computer algorithm to recognize subtle changes in cells when treated with certain experimental compounds. The computer ‘neural network’ predicted almost 100% the results for cells treated with 100 mysterious compounds, even at the various level of dosage.”
Bulgaru continues, “Novartis’ machine learning algorithms are able to classify compounds with the same visual effects on particular cells – and do it with incredible speeds. That means drug discovery insights that would otherwise take months, if not years, for humans to generate will only take a matter of minutes (if not seconds!).”
Given the skyrocketing costs of drug development, there is a particular need to bring this kind of acceleration to clinical research—and machine learning is an important tool for meeting this need. In a roundup of innovations that can reduce time and cut costs in clinical research, PharmaTimes.com showcased several that involve machine learning, from the use of data-driven analytics to improve trial design and planning to the use of AI and machine learning to improve productivity in pharmacovigilance.
Predictive analytics for the improvement of risk monitoring is another innovation that PharmaTimes.com highlighted, noting “When sponsors leverage artificial intelligence (AI) and machine learning to analyze multiple non-identified global healthcare data sets, they can make evidence-based predictions to improve decisions at every stage in the development and commercialization process.”
The wide-ranging potential for machine learning in pharma R&D is huge. Strides are already being made, but perhaps the most important need for AI is the “fuel” to fire that potential—and that is data. “I think actually the biggest challenge, the biggest need for us is really the high-quality data sets,” said Morten Sogaard, VP and head of Target Sciences at Pfizer, in an interview with Bio-IT World. “And then of course any company would need—it doesn’t necessarily have to be a whole army—but need to have a few truly brilliant data scientists in order to execute.”
Matt Clark echoes this, citing the ‘garbage in garbage out principle’ and explaining that algorithms can only extrapolate from what is already known. “The future of the research is a hybrid model where machines augment and enhance the work being done by human experts,” he says. “Deep learning will be at the heart of this change, yet it isn’t a magic bullet. Just like any good human researcher, these virtual researchers require quality data with which to work. Therefore care must be taken to ensure that the input data is curated and annotated to make sure that the end result is meaningful”
Frederik van den Broek
In Elsevier’s Professional Services team, Frederik is the European lead consultant on data integration and analytics projects throughout the life science, chemistry and engineering domains using commercial, proprietary, and public data sources. Before joining Elsevier in late 2014, Frederik van den Broek spent 10 years in R&D IT consultancy at Tessella, working for R&D organisations of multinational companies in a variety of industries: life sciences, consumer products, energy, chemical and petrochemical and hydraulic engineering. He has a broad international experience and a deep understanding of science and technology. Frederik holds a doctorate in Chemical Physics from the University of Amsterdam / FOM Institute AMOLF and a master’s degree in Chemistry from Utrecht University.
R&D Solutions for Pharma & Life SciencesWe’re happy to discuss your needs and show you how Elsevier’s Solution can help.