Bioinformatics

A core resource for the CTSI community, the Bioinformatics Program helps translational and clinical researchers at BU apply biomedical informatics methods and tools to their research. The Bioinformatics Division has two principal components—a Translational Bioinformatics component, led by Avrum Spira, MD, MSc, and a Clinical Research Informatics component, led by William Adams, MD. The program’s design is user-friendly and interfaces with NIH and other institutions’ systems thanks to the open source, standards-based informatics for Integrating Biology and the Bedside (i2b2) platform.

The Bioinformatics Program aims to support informatics across the translational continuum by:

  • Developing, demonstrating, and disseminating new solutions
  • Collaborating with stakeholders in the CTSI hub community and across the CTSA network
  • Adopting informatics best practices from other centers

Translational Bioinformatics

The Translational Bioinformatics Program helps researchers apply bioinformatics methods to their research. Our faculty strives to improve the diagnosis, prognosis, and treatment of human disease by applying computational approaches that leverage high-throughput molecular data.

The goals of the program are to:

  • Offer expertise in applying bioinformatics to translational research.
  • Establish a facility devoted to translational bioinformatics resources that supports the computational needs of researchers.
  • Develop a translational bioinformatics education and training program that enables scientists in the biological sciences to apply computational tools to advance research.

The Translational Bioinformatics group includes:

openSESAME (Search of Expression Signatures Across Many Experiments)

CTSI developed an approach to identify relationships between datasets based on patterns of gene coexpression, the so-called “Search of Expression Signatures Across Many Experiments,” and created openSESAME to allow the scientific community to apply this approach to what is currently about 75,000 Affymetrix human gene expression profiles obtained from the Gene Expression Omnibus (GEO) at the National Center for Biotechnology Information (NCBI). An investigator who derives a gene expression signature (sets of genes that are expected to be coordinately induced or repressed) in one setting can use this signature as input to openSESAME, which ranks the compendium of available datasets by the degree to which these genes are similarly coordinately differentially expressed and provides tabular and graphical summaries of these results for further exploration. This approach can potentially identify shared molecular drivers between seemingly disparate biological conditions and can identify unexpected new therapies for disease including repurposing existing FDA-approved therapies.

The openSESAME resource is being expanded to include all raw data from GEO that was generated using Affymetrix microarray platforms in samples from humans or from several mammalian model organisms, resulting in a resource with far greater potential for identifying conditions that give rise to shared gene coexpression. In addition, methods are in development to facilitate the identification of experimental variables that are associated with the regulation of a gene expression signature in a given dataset.

One of the most exciting applications of the openSESAME approach is in repurposing existing therapeutics for alternative uses. The process of drug repositioning can be performed in silico through a process similar to online dating: gene expression profiles of cells treated with various compounds or of clinical samples from cases and controls are used to construct signatures of drugs and diseases, respectively, and all possible drug-disease pairs are automatically ranked, with the openSESAME algorithm playing the “matchmaker.” In this case, “opposites attract”: if the signature of a given drug is the mirror image of a given disease signature, there is reason to believe that this drug may be able to reverse biological processes set in motion by that disease. A key advantage of this approach is that, since it is performed in silico, it is extremely rapid and inexpensive once the underlying gene expression data has been collected. This kind of approach has been previously employed by CTSI, in collaboration with investigators in the Division of Computational Biomedicine at BUSM, to identify existing drugs that might serve as new therapeutics treatments.