What's the score?
"Proteomics has tremendously advanced, even in the few years since I arrived at Lunenfeld," Gingras says, "and it has been very exciting to watch this progress. But new possibilities also bring up new challenges, things like ensuring that data quality is kept high and that the data can be shared in an appropriate format with the whole scientific community."
Gingras believes specialized journals like Molecular and Cellular Proteomics, of which she is an editorial board member, have been pioneers in this respect, but more general journals still need to improve their mechanisms to ensure data quality and data sharing. "Funding agencies should also step up and help facilitate funding of projects that address these issues, such as maintaining data repositories."
Of the many challenges that fall under the umbrella of figuring out what to do with all the data, Gingras has most keenly been intrigued with data scoring.
STRIPAK and the Golgi. Another of Gingras' recent discoveries with STRIPAK is its role in Golgi orientataion. Here, cells depleted of STRIPAK component CCM3 by esiRNA (left) or chemical siRNA (right) show reduced numbers of Golgi properly positioned (white asterisks) in response to a wound (shown by dashed line). esiLuc and siScramble are negative controls. From J. Biol. Chem. 286, 25065-25075.
"It is a major issue in the field," she says. "Proteomics can identify all sorts of protein-protein interactions, but how can we be sure they are genuine? I hate to imagine a proteomics study producing a novel interaction, but later on a poor graduate loses weeks or months of work unsuccessfully doing biochemical follow-ups because the interaction was spurious."
To rescue frustrated students across the world, Gingras teamed up with colleague Alexey Nesvizhskii at the University of Michigan to develop an innovative computer program called SAINT (for Significance Analysis of Interactome); this groundbreaking probability-based program analyzes several metrics to quantify protein interactions in a data set as true or false.
The seeds for SAINT go back more than eight years, starting with some informal discussions and idea exchanges between Gingras and Nesvizhskii when they were both postdocs with Aebersold. "Back then, though, we didn't have the technology to realize our ideas, but a couple of years ago we reconnected and thought, you know, this could work now."
Gingras believes tools like SAINT will have tremendous impact on how researchers analyze, score and report their protein-protein interactions. With Mike Tyers at the University of Edinburgh in Scotland, Gingras has been working on a complementary program called ProHits that lets scientists store, search and analyze mass-spectrometry data. She now is working on interfacing ProHits and SAINTS to create a thorough software tool.