Sage Bionetworks Research is based on its outstanding portfolio of coherent datasets obtained through partnerships with academic and pharma collaborators.

Sage Bionetworks scientists work with academic and commercial partners on comprehensive molecular and clinical datasets to create validated disease models that improve the speed and efficiency of therapeutic drug development. Details of recent success can be found in the appended publications.


The Sage Bionetworks team of Network Biologists, Systems Biologists, Statistical Geneticists and Computational Biologists has continued the track record of leadership and innovation begun at the Merck & Co., Inc. subsidiary Rosetta Inpharmatics. They currently work with several large pharmaceutical corporations and biotechnology companies as well as a range of international academic partners. The research projects benefit both the specific collaborators and the larger scientific community because the results will also be accessible in the Sage Bionetworks Commons one year after the conclusion of the research projects.

The Sage Bionetworks projects span cardiovascular disease, cancer, metabolic disease and neurobiology. The resultant analyses and predictive models are used for six primary purposes:

  1. Connecting associated SNPs to true gene underlying mechanism via Genetics of Gene Expression: Using GWAS and other association between DNA variation and a clinical phenotype to understand what genes and ultimately mechanism underlie that association. Sage Bionetworks uses its human eSNPs, SNP-set-enrichment, mouse causal genes, and similarities between human and mouse networks to determine plausible genes and network neighborhoods through which the information encoded in that DNA variation manifests as phenotype.
  2. Identifying new targets and progress through validation as disease genes toward pharmacologic validation: Predicting genes that contribute to disease phenotypes using causality and network modeling. Using multiple examples that validate based on a single-gene intervention in a model system, and ultimately progressing toward in vivo pharmacology.
  3. Repositioning drugs: This is a special facet of new target identification, where the project starts with a number of targets for which good, "safe" compounds exist, and then applying the full repertoire of Sage Bionetworks computational and standard laboratory approaches to validate the target and test the compound for an indication in preclinical species or humans.
  4. Terminating drug development programs with confidence that opportunities to segment the target population were fully explored: Taking a Phase II or III trial where efficacy is not seeming strong, or where adverse experiences appear mechanism-based and using genetics in the trial plus the network approaches outlined in first example to demonstrate that a significant segment of the population for which the drug would have substantial net benefit is unlikely to exist.
  5. Defining clinically relevant subpopulations: Similar to above, but typically starting at an earlier stage to incorporate hypotheses about population segments early enough in the development process that they are easily tested prospectively.
  6. Avoiding drug liabilities: Applying a pipeline of standard checks to expression profiling from knockout, siRNA and compound treatments for a target that encompasses mapping the expression signatures to all relevant tissue networks, looking to see what annotations and other gene expression signatures map to the modules where those intervention signatures map, and following up any leads.


^ Back to top