Skip to navigation Skip to main content Skip to footer

Approved Research

Multi-task Learning for Cancer Phenotypes

Principal Investigator: Professor Giovanni Parmigiani
Approved Research ID: 94470
Approval date: March 9th 2023

Lay summary

Decisions about cancer prevention and early detection can be complicated and often boil down to whether the odds of getting cancer, for you, are big enough to justify the trouble, cost and potential adverse effects of what your doctors will need to do to catch your cancer early, if it happens. The more accurate these odds are, the better the decision you will make. Cancers do run in families, and families who are prone to cancer can experience cancer in different organs. For example many relatives of ovarian cancer patients experience breast cancer. This is because there are some fundamental genetic mechanisms that, when defective, can cause trouble on multliple body parts. What we want to do here is to use this concept, and some advanced computational method we developed, to improve the calculation of the odds of getting cancer based on one's genes.