Pharma R&D Today
Ideas and Insight supporting all stages of Drug Discovery & Development
How Much Can Text-Mining Tools Achieve?
Posted on July 25th, 2016 by Patricia Morgan in Chemistry
Text mining has become an essential tool in the arsenal of disease biologists, medicinal chemists and other drug developers. There simply isn’t the time for a researcher to read all the relevant published literature to find answers—in particular when the goal is to retrieve an exhaustive set of information or precise data on a particular drug. The volume of publications would daunt even the most avid reader!
An answer to the challenge comes in the form of text-mining tools, which use software that can actually read published abstracts or full-text literature. With the right solution, a researcher can create a query and let the tool do the time-consuming work of finding every statement, phrase or data point that matches it. Then the researcher just has to work with the filtered results for the specific query. Results are more comprehensive than any manual search strategy and often provide valuable insights into testable hypotheses.
Want to identify all the possible biomarkers reported for a particular medical condition? Or rapidly retrieve the published toxicity, efficacy, pharmacokinetic, safety and metabolic data for a particular drug candidate? Or explore publications about other drugs being tested for similar purposes? Text mining of publicly available databases is currently the only rapid way to tackle such questions. These approaches can also be customized so that the text mining process can include internal databases, making it possible for a researcher to capture a broader range of material in a single search.
It sounds great but anyone who’s struggled with technology not working quite as well as hoped could be forgiven for being skeptical. Can software really perform that well, especially when it comes to something as fundamentally human as reading?
The answer might surprise you. The accuracy of the leading text-mining software is nearing that of manual efforts, with some solutions able to reach 85% of the accuracy of a human in just a fraction of the time.
The best text-mining solutions can—and indeed already do—provide answers to many of the complex questions involved in drug development. While nothing will replace the researchers who actually interpret the data, make the critical decisions and develop drugs that save lives, the right text-mining solution gives the whole research process a solid foundation.
This article goes into more detail on the potential of modern text-mining solutions, looking at how users can leverage the information and the potential for the whole drug development workflow.
All opinions shared in this post are the author’s own.
R&D Solutions for Pharma & Life SciencesWe're happy to discuss your needs and show you how Elsevier's Solution can help.
Product Marketing Manager
- A Hackathon Demonstrates the Need for Deep Learning in R&D
- The Importance of the Team in Drug Development
- Where and How to Find Co-Occurrences in Disease Biology Research
- Text Mining for Pharmacovigilance Purposes
- The Fundamentals of Good Data Stewardship