Pharma R&D Today
Ideas and Insight supporting all stages of Drug Discovery & Development
A New Approach to Drug Analytics
Posted on December 18th, 2015 by Neal Katz in Pharma R&D
Technological advances in the retrieval and assessment of pathway and bioactivity data are giving rise to new in silico profiling and modeling methods. Our new white paper, Combining Pathway and Bioactivity Data for New Insights, examines a straightforward and elegant means to create predictive models of compound behavior and describes some of its applications.
Read this excerpt on how to use predictive modeling to generate new target ideas.
Generating new target ideas for potential drugs
Predictive modeling is equally effective in the generation of target ideas for compounds that have not yet entered the drug discovery process. In the earliest stages of drug discovery with a new test compound, there is not the same wealth of data about its effects. Comparative studies become more important in such cases as it is the behavior of similar compounds that will help to predict the in vivo effects of the candidate. This type of predictive modeling is useful both in the determination of a test compound’s efficacy against a single disease of interest and in the discovery of target ideas that may otherwise have been overlooked.
Generating target ideas for new test compounds:
- Use text mining to identify all the targets relevant for diseases of interest
- Find existing drug-like compounds that bind tightly to those targets
- Rank the compounds by number of targets affected and affinity of binding
- Select compounds for further exploration and testing
This method uses the same set of techniques as the predictive modeling for drug repurposing: comprehensive text mining based on natural language search capabilities; retrieval of bioactivity data from an appropriate database, preferably one with normalized affinity values; and ranking the results based on number of targets and affinity. The difference lies in the subject of each stage of the study.
This process helps explore new mechanisms of action and provides an information-based method for initial steps of target validation. Text mining is applied to identify all of the targets that are relevant for each disease of interest based on sequencing data about the relevant pathways. Once the list of targets is known, compounds that bind to one or more of the targets are identified and the structure–activity relationship for each interaction is retrieved. This information allows the ranking of the compounds by the number of targets and their affinities.
The more disease targets a given compound binds, the greater the likelihood of the compound affecting the disease, although it should be noted that a single high-affinity target may be sufficient. It is also possible to look at off-target binding at this point to see if there are adverse events that would preclude the use of the drug in this manner.
When the known compounds have been identified and ranked, a thorough study comparing the test compound and known compounds will reveal the best target indications for the new compound based on structural similarities. This provides a solid foundation for further in silico and laboratory studies.
Read the full white paper, Combining Pathway and Bioactivity Data for New Insights.
R&D Solutions for Pharma & Life SciencesWe're happy to discuss your needs and show you how Elsevier's Solution can help.
Director of Solutions Marketing
- Navigating the Virus Regulation Pathway through Text Mining and Knowledge Graph
- Combining Data to Find New Indications, Develop New Targets & Predict Adverse Events
- The Fundamentals of Good Data Stewardship
- Sources of Innovation for Drugs of the Future