Pharma R&D Today
Ideas and Insight supporting all stages of Drug Discovery & Development
Building models for rare diseases
Posted on June 2nd, 2016 by Maria Shkrob in Pharma R&D
We continue the series of stories about the exciting collaboration between the UK charity Findacure and Elsevier. Findacure’s mission is to empower rare disease patient groups, facilitate treatment development, and to campaign for a receptive research environment, and we are proud to be a part of this mission. Using multiple Elsevier resources, such as Elsevier Text Mining, Pathway Studio, Reaxys Medicinal Chemistry, PharmaPendium, Scopus, ScienceDirect, and Mendeley, we help creating a comprehensive view of disease, identifying repurposing candidates, and building a scientific community. Read our previous messages:
- Introduction: Drug repurposing for rare diseases
- Using text Mining to Find Treatments for Rare Diseases
Last time we used text mining to identify proteins, small molecules, diseases, and cell processes related to CHI. We collected the dots, and now it is time to connect them – create a comprehensive disease model for CHI in Pathway Studio.
Elsevier’s Pathway Studio is a web-based software tool that allows you to navigate 6.2 million relations between biological concepts (such as proteins, small molecules, cells, diseases, and cell processes) extracted from extensive corpus of Elsevier and non-Elsevier publications (3.5 million abstracts, and 25 million full text publications) using text mining. In Pathway Studio literature-derived information is combined with annotations from other data sources, collection of pathways, and enhanced with algorithms for visualization, summarization, filtering, and network analysis.
In this project we used Pathway Studio to summarize the published research about CHI, to build a pathway for CHI and one of the drugs that are being tested against it (sirolimus).
Building a CHI disease model in Pathway Studio
Knowing the key players and their relation to the disease of interest is already useful, as it provides insights into potential targets and biomarkers for the disease; the next level of structuring the knowledge from the literature is to build a pathway, a model of the disease, – position key players and interactions to represent the order of the events. Understanding the bigger picture is crucial, for example to predict and explain drug sensitivity: if a dysfunctional protein causing the disease acts downstream from a drug target, a drug might be ineffective.
While over 100 diseases already have pre-built models in Pathway Studio, CHI wasn’t among them, as it is an ultra rare condition, and yet we didn’t have to build the CHI model from scratch, as Pathway Studio contains over 1800 manually curated pathways illustrating:
- Metabolic processes
- Disease mechanisms
- Toxicity mechanisms
- Pain mechanisms
- Immunity and inflammation
These can be used as building blocks to create new pathways. For example, CHI (which is deciphered as congenital hyperinsulinism) is a disease characterized with abnormally high levels of insulin in the blood, allowing us to use canonical pathways for insulin production and release from Pathway Studio collection as a starting point to build a pathway for CHI.
At first we highlighted the members of the canonic pathway that were studied in the context of CHI. Then using the information from CHI literature we added proteins with understood connection to the CHI to the scheme. Next we added information about targets and treatments for CHI. To finalize the pathway, we removed proteins that were not mentioned in CHI literature from the original pathway to make the final scheme more comprehensive.
Our collaborators from Findacure appreciated the interactive model – a visual summary with annotations for molecules and connections, and are interested in continuing this project and building an additional, very light model of CHI, the one they could share with parents of children diagnosed with CHI to explain how a particular mutation leads to the symptoms, and why a particular treatment option would be the most effective.
Using text-mining engine we identified over 60 genes/proteins connected to CHI. Some of them are well-established as contributors to CHI mechanism (these proteins were added to CHI pathway), the role of others still has to be elucidated. Using Pathway Studio’s database of interactions extracted from literature, combined with Gene Ontology and Elsevier ontology, we were able to annotate all potential players with their function, and connect them to insulin-related processes, such as insulin production and release. The results will be shared with researchers collaborating with Findacure to provide additional insights and help generate hypotheses about these proteins role in CHI mechanism.
Having an up to date disease summary is crucial to create effective treatments and suggest biomarkers. In the next two years we will keep monitoring newly published research about congenital hyperinsulinism to expand the disease pathway with new players and connections.
- Learn more about Pathway Studio and sign up for a free trial
- Learn more about Pathway Studio pathways collection: “The Pathways to Understanding Diseases. Deciphering complex biological processes”
- Take a look at sample pathways (click on a tab “Sample Pathways” in the left top corner)
- Listen to a webinar “Mobilizing Informational Resources for Rare Disease” where we share the story of Findacure and Elsevier collaboration in details
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.
Sr Bioinformatics Scientist
- The Future of Chemistry R&D Will Depend on Deep Learning
- When Looking to Accelerate the Pace of Research, Having the Right Multidisciplinary Tools Matters
- To streamline early discovery, a cross-disciplinary, data-driven workflow process is key
- A Hackathon Demonstrates the Need for Deep Learning in R&D
- Quantitative Proteomics: A Major Piece in the Precision Medicine Puzzle