Writing the Genetic Instruction Manual

Roth, a professor at the University of Toronto’s Donnelly Centre and Senior Scientist at Mount Sinai Hospital’s Lunenfeld-Tanenbaum Research Institute, belongs to an international team of researchers who are mapping human protein interactions. The group, which is co-led by Marc Vidal at the Dana-Farber Cancer Institute and Harvard Medical School, published a paper in Cell today describing a systematic map of how 13,000 proteins, each produced from the plans of a different gene, connect to one another.

Just as a mechanic needs more than a list of parts to fix a car, Roth says the human genome doesn’t give scientists all of the information they need to understand how the parts of a cell connect.

“One of our major conclusions is that proteins involved in cancer are more likely to connect to each other than they are to connect to other types of protein. This means if you see a protein that isn’t known to be linked to cancer in a cluster with cancer proteins, we should be suspicious that it may also be a cancer protein,” says Roth.

This increased understanding of protein interactions could lead to better avenues for targeted therapies. The study focused on cancer, but the map will help scientists better understand a variety of diseases involving protein interactions. Researchers are just beginning to tap what can be learned from the network of molecular and genetic interactions in the human body.

“It’s a big step forward in a long walk,” says Roth, a Canada Research Excellence Chair in Integrative Biology affiliated with the Departments of Molecular Genetics and Computer Science at University of Toronto and a Senior Fellow of the Canadian Institute for Advanced Research. “But we’re starting to see how the pieces encoded in the human genome physically come together.”

The study also found that most of what researchers knew about protein interactions involved proteins that were already well studied. Roth and his collaborators want to ensure the unstudied proteins aren’t forgotten.

“When we look systematically across all proteins, including the unpopular proteins — the ones that aren’t as well studied — we see almost as many protein interactions among the unpopular proteins as among the popular proteins. That means there’s a lot more to this system than scientists first realized,” says Roth.

An example is the gene MAPK1IP1L, which hasn’t been extensively studied or reported as a cancer protein in humans. But, three of its 13 interaction partners are known cancer genes. Also, disruption of this gene is associated with tumor formation in mice, so it is probably worth a closer look.

The team estimates that no more than five-to-ten percent of the direct connections between proteins are represented on the map. The next phase of the study will expand the project to one protein from each of 17,000 genes. Many genes provide plans for additional alternative proteins, often with different interactions.  The ultimate goal is to map interactions among all proteins from the 20,000 genes that encode proteins in the human body.