Bioinformatics analysis in the style of social network analysis yields maps of gene networks for about 400 different human cell and tissue types
Bioinformatics analysis in the style of social network analysis yields maps of gene networks for about 400 different human cell and tissue types

Thousands of genetic variants have been associated with different diseases, but these variants still retain their most treacherous secrets—the mechanisms by which they contribute to disease processes. The trick is these mechanisms typically involve multiple conspirators—not just genes, but also transcription factors, enhancers, and promoters. If these networks could be exposed, it would be possible to develop better diagnostic tests and personalized treatments for patients.

This possibility motivated researchers based at the Swiss Institute of Bioinformatics to adopt a network-centric view to the analysis of genome-wide association studies (GWAS). More to the point, these researchers took inspiration from the kind of analysis that is used in the world of social networking. Instead of looking at interconnections to learn about social network users, the researchers tried looking at molecular regulatory pathways to learn about the genetic “influencers” of disease.

Innovative software tools, such as the Pathway Scoring Algorithm (Pascal), allowed the scientists to construct accurate “maps” of gene networks for about 400 different human cell and tissue types, ranging from immune cells to brain tissues, whereas previous studies were limited to just one or few tissues. Each of these networks describes hundreds of thousands of regulatory interactions among thousands of genes, giving the first global view of the “control system” of diverse cells and tissues.

The details of this work appeared in two recent articles, one of which appeared January 25 in PLOS Computational Biology (“Fast and Rigorous Computation of Gene and Pathway Scores from SNP-Based Summary Statistics”), and one of which appeared March 7 in Nature Methods (“Tissue-Specific Regulatory Circuits Reveal Variable Modular Perturbations across Complex Diseases”). These articles describe how the Swiss researchers, in collaboration with researchers from the Broad Institute of MIT and Harvard, found that genetic variants disrupt components of regulatory networks in disease-specific tissues, giving new insights on disease mechanisms that may lead to targeted treatments that are more effective and have fewer side effects for the patient.

“Pascal was designed to be fast, accurate and to have high power to detect relevant pathways,” the authors of the PLOS Computational Biology article wrote. “We extensively tested our approach on a large collection of real GWAS association results and saw better discovery of confirmed pathways than with other popular methods.”

In the Nature Methods article, the researchers went still further. They undertook the mapping of perturbed molecular circuits that underlie complex diseases.

“The challenge is that over 90% of disease variants lie outside of genes, in regions of the genome that are still poorly understood” said Daniel Marbach, Ph.D., one of the researchers who is affiliated with both the University of Lausanne and the Swiss Institute of Bioinformatics. “These regions can have regulatory functions, which are sometimes disrupted by genetic variants. Things get even more complicated as the regulatory relationships may vary between different tissue types. For example, a certain gene may activate another one in the liver, but not in the heart.”

The researchers combined data from an international research consortium (Functional Annotation of the Mammalian Genome, or FANTOM) and novel analysis techniques. This approach allowed the researchers to create the largest collection of such networks to date, describing the regulatory interactions among over 19,000 genes.

“We developed a comprehensive resource of 394 cell type- and tissue-specific gene regulatory networks for humans,” the authors wrote. “Integration with 37 genome-wide association studies (GWASs) showed that disease-associated genetic variants—including variants that do not reach genome-wide significance—often perturb regulatory modules that are highly specific to disease-relevant cell types or tissues.”

The authors added that they believed that the resource they developed opens the door to systematic analysis of regulatory programs across hundreds of human cell types and tissues.

In a large study including genetic data for diverse neurodegenerative, psychiatric, immune-related, cardiovascular, and metabolic disorders, the researchers found that disease variants often affect groups of genes that were densely interconnected within regulatory networks, confirming their hypothesis. Moreover, these affected network components pinpointed with remarkable precision cell types or tissues that are implicated in disease processes.

“For example, people with schizophrenia were found to have genetic variants that perturb interacting genes in brain tissues that are responsible for cognitive and emotional behavior,” explained Dr. Marbach. “Genetic variants associated with obesity impact genes that interact in tissues of the intestinal system.”

“Our work shows that accurate maps of gene networks for different tissues will be of tremendous value to advance our understanding of how diseases start and progress,” added Sven Bergmann, Ph.D., a senior author of both studies. “[Such an understanding] is essential to design targeted treatments and to identify patient groups that respond to these treatments in a personalized medicine setting.”

Also of Interest