Connecting the dots of our genome

One of the central questions in human biology is to understand how our genes determine which diseases we get and how severe they might be. Knowing just the DNA sequence, or the blueprint, is not enough. We must figure out how proteins, the genes’ products, work too.

Now an international team of researchers, jointly led by Dr. Fritz Roth (at Mount Sinai Hospital’s Lunenfeld-Tanenbaum Research Institute and the Donnelly Centre of the University of Toronto), and Dr. Marc Vidal (with the Dana-Farber Cancer Institute and Harvard Medical School in Boston), have produced the largest ever map of human protein interactions. This publicly available resource will be invaluable to anyone trying to understand complex genetic traits and develop new disease therapies.

“It is realistic to think that many of the people reading this will have their genomes sequenced within their lifetimes. The next challenge is to figure out what their genomes mean,” says Dr. Roth. “You cannot figure out how the car works based on the parts list. You have to know how they fit together.”

This is because genes do not do the work in a cell. Rather, the work is usually done by the proteins that genes provide the plans for. Some pairs of proteins stick together, or ‘interact’ when they are in close contact with each other. These interactions underlie all of cell’s biology and mediate processes such as gene expression, cell metabolism, and transporting other molecules within a cell.

Having a detailed map of protein interactions bring us one step closer to understanding the relationship between our genes (genotype) and our physiology in health and disease (phenotype).

Drs. Roth and Vidal, and their colleagues, analysed direct interactions in pairwise combinations between 13,000 proteins. Out of 85 million possible interactions they found 14,000 directly-interacting protein pairs. This more than doubles the previous set of known interactions, making it the largest ever experimentally determined human protein interaction map.

“We’ve managed to peer into the car and connect a fraction of the parts,” says Dr. Roth.

The study reveals several important findings. The new map can be used to identify novel genes involved in diseases. If a novel protein, which we know nothing about, interacts with a known protein that has a role in a disease, then the novel protein is highly likely to be involved in that same disease. Dr. Roth and colleagues illustrate this point by identifying a novel cancer gene STAT3 based on its interactions with known cancer genes. Their finding was confirmed when STAT3 subsequently became included into the cancer gene database based on independent evidence.
 Further unbiased analyses identified 100 strong cancer candidate genes, 60 of which were connected to known cancer molecular pathways. Some of these genes are completely novel. This shows the potential of the human interaction map in revealing new disease genes and promising therapeutic targets.

Mapping all human protein interactions is a colossal task and will require several different approaches. This is because not every method can find every protein interaction. Dr. Roth estimates that they found, using a yeast two hybrid method, 5-10% of all protein interactions, a substantial increase from their previous paper that reported 1% of interactions.

“Although much sweat and some tears were put into analysing this new map, it is clear that we have only scratched the surface of what these interactions can tell us about human disease. It is personally very exciting to anticipate the discoveries to come, as it passes from our hands into the research community,” says Dr. Roth. Lunenfeld-Tanenbaum Research Institute