Dr. Neta Zach, Chief Scientific Officer of Prize4Life-Israel, recently presented a talk about the PRO-ACT (Pooled Resource Open-access ALS Clinical Trials) platform, which houses the largest ALS clinical trials dataset developed to date, at the American Medical Informatics Association, Joint Summits on Translational Science. The conference, held April 7-11 in San Francisco, California, brought together attendees to discuss advances in bioinformatics and medical informatics. Dr. Zach reports on a few particularly exciting sessions with relevance to drug discovery and clinical information in neurodegenerative disease.
Translational Bioinformatics Opening Session and Keynote Presentation: Human Genetics and Translational Bioinformatics to Guide Drug Discovery
Robert Plenge, Vice President, Merck Research Laboratories (MRL) and Worldwide Head, Genetics and Pharmacogenomics (GpGx) delivered the Keynote presentation at the Translational Bioinformatics Opening Session on Human Genetics and Translational Bioinformatics to Guide Drug Discovery. He underscored one of the primary obstacles for current drug development: the high degree of uncertainty regarding probability of success in the human trials, as many treatments found safe and effective in preclinical studies ultimately end in failed clinical trials due to toxicity, lack of efficacy, or failure to improve on the standard of care. Can human genetics and translational bioinformatics improve the success rate of human trials? Dr. Plenge posited that human genetics can provide insight into the optimal pathways for drug targeting and the effect of genetic manipulations. These predictions can be even more precise now, with the availability of Genome Wide Association Studies (GWAS) data.
What are the best strategies to leverage human genetics to optimize drug discovery?
He proposed two approaches:
The first involves broadly surveying genes implicated in a disease to identify critical pathways for therapeutic targeting. A successful example from rheumatoid arthritis (RA) was the identification of the CD40 pathway as the convergent pathway affected by mutations in over 100 loci. The importance of CD40 signaling in RA became apparent by applying statistical analyses to variable loci in the GWAS data and cross-referencing this information with gene expression analysis of mouse models of RA. Similar approaches can be used in ALS using the ALSgene and ALSoD databases.
The second approach involves focusing on a single gene with a direct relationship to the disease phenotype. By surveying large genetic data sets, rare alleles of this gene can be identified and their function characterized. One such example comes from the area of heart disease, where alleles associated with lower LDL were shown to be protective. Drugs mimicking these alleles are now in Phase II clinical trials.
The large data sets needed to perform this type of analyses are becoming increasingly available to the broad research community with efforts to integrate and analyze millions of patient electronic health records from hospitals and clinics. Neurobank (read more about it here) is a promising initiative of this kind in the area of ALS.
Panel- Dissemination of Pharmacogenomic Knowledge: Establishing a Pathway to Support Clinical Implementation
Panelists: James Hoffman, St. Jude Children’s Research Hospital; Michelle Whirl-Carrillo, Stanford University; Robert Freimuth, Mayo Clinic; Josh Peterson, Vanderbilt University.
Pharmacogenomics is the analysis of how a patient’s genetic makeup determines their drug response. Pharmacogenomic research is a burgeoning field with direct clinical applications, which has lead to establishment of several large initiatives to support integration and dissemination of findings from pharmacogenomic research.
As the first stage, all of the relevant open access data will be housed in a single database called the pharmacogenomic knowledge base or PharmaGKB.
During the second stage, the data will be converted by a specialized committee of researchers and clinicians into a set of clear tables with guidelines for clinical interpretation (for example, a patient with these genetic results will respond best to medication X with 90% probability). These guidelines will be developed by the Clinical Pharmacogenetics Implementation Consortium or CPIC, a shared project between PharmGKB and the Pharmacogenomics Research Network. The basic questions being addressed by CPIC are what genotypes must be taken into account in clinical care for a given disease and how to overcome barriers to implementation of pharmacogenetic tests into clinical practice. This has already proven to be of utmost importance in cancer care, where some genetic alleles make certain drugs toxic and others ineffective.
As a third stage, the Pharmacogenomics Research Network (PGRN) and the Translational Pharmacogenomics Programs (TPP) included in the PGRN will be responsible for incorporating the CPIC’s data into the electronic health records in order for clinicians to have direct access to the information.
Another noteworthy session was the Clinical Informatics Research Opening Session and Keynote Presentation by Richard Platt, Professor and Chair of the Department of Population Medicine and Executive Director of the Harvard Pilgrim Health Care Institute, discussing “can you answer clinical trial types of questions using electronic health records?”. One ambitious initiative to address this question is the Patient-Centered Outcomes Research Institute (PCORI)’s PCORnet, a newly established consortium that will use electronic health data to conduct comparative effectiveness research in disease areas including ALS. The PCORnet initiative aims to enable integration of large data sets from diverse sources, including pharmacies, hospitals and clinics, in sufficient detail to provide the necessary data package for the FDA. This mission is especially challenging due to the substantial number of confounding factors across different studies, such as differences in inclusion criteria variability in the extent of longitudinal information available for different data types.
Over the last few years, the clinical research field has seen an explosion of data from electronic health records, hospitals, pharmacies, supermarkets and insurance claims. Each research question requires an individualized plan for how to integrate and analyze the diverse data sources to obtain an insightful result. This is the time to start gearing up with new research ideas! – Neta Zach.