Medical records may be private information, but a patient’s genes belong to the entire family. And that raises a quandary for medical researchers: When they find something troubling in a subject’s genome, should they pass it along to the family? Even after the patient dies?

Dozens of the nation’s top researchers, bioethicists and federal health officials gathered Thursday at the University of Minnesota in an effort to answer those questions and offer guidelines for ethics dilemmas created by the burgeoning field of genomic research.

In a recent survey, cancer patients were asked to consider the case of a hypothetical patient named “Pat” who was diagnosed with pancreatic cancer in 2009 and enrolled in a research project at the Mayo Clinic. Three years later, researchers running tests on Pat’s blood discovered a mutation with strong links to breast cancer. Should they inform Pat’s 22-year-old daughter so she can get regular cancer screening? Should other relatives be told that Pat carried a gene now linked to pancreatic cancer? What if Pat wanted the research results kept private, but has since died?

Breakthroughs in genomic testing allow researchers to uncover risks with serious health implications for a subject’s biological relatives, but U.S. law and research rules assiduously protect the privacy of research subjects, said Susan Wolf, founding chair of the U’s Consortium on Law and Values in Health, Environment & the Life Sciences, which hosted the conference.

In addition, genomic research often carries on for years — long after individual study participants have died. Relatives are curious, but the federal Health Insurance Portability and Accountability Act (HIPAA) preserves the privacy of an individual’s medical and laboratory records for 50 years after their death, though it allows some exceptions.

“So there is this tension,” Wolf said. “We have to anticipate, what are we going to do with all of this information after the death of a participant?”

Wolf is one of three principal investigators among the U, the Mayo Clinic and the University of California, San Francisco (UCSF) who are spearheading a $2.4 million project funded by the National Institutes of Health to guide the disclosure of “incidental findings” in genomic research.

Wolf and her 15 collaborators presented their preliminary recommendations Thursday at the U. They plan to produce a final report to be published next year in the Journal of Law, Medicine & Ethics.

For now, the group argued that researchers should be allowed — but not required — to make “passive disclosures” to family members who seek out the results of genetic tests on their kin. Researchers wouldn’t be obligated to track down relatives, however, because of the huge potential expense.

They also recommend that participants enrolling in studies be asked about their preferences for disclosure, including after they die.

Even so, they say researchers should be allowed to override those preferences and seek out relatives if a genetic mutation is strongly linked to serious diseases and disclosure to family members is likely to prevent serious harm.

Wolf said genetic information should be turned over to relatives only if it has been shown to be valid, reveals a substantial health risk, offers net benefit to the family and comports with the law. Researchers should generally avoid sharing a participant’s full genetic sequence without the participant’s expressed consent, Wolf said.

Some research shows a disconnect between current research practices and what patients expect. Barbara Koenig, a biomedical ethicist and anthropologist at UCSF, and Carmen Radecki Breitkopf of the Mayo Clinic, surveyed participants in Mayo’s pancreatic cancer biobank and the public. Fully 99 percent said they expected to be told if researchers found something harmful, and 85 percent said they expected to be told if something beneficial was discovered in their genes.

Only a third thought researchers should keep information from family members about a harmful mutation strongly associated with breast cancer if the research participant were still alive.

Sixty percent of the participants who were interviewed said genetic information belongs to the whole family, Koenig said. Nearly 77 percent said it’s a researcher’s responsibility to offer the information whether the family members ask for it or not.

“They didn’t like the idea that these researchers would say it’s just not my job to do that,” Koenig said.

But that may be impractical. The Broad Institute at Harvard University just put a database online of exome data from nearly 63,000 unrelated individuals so that a coalition of researchers could study them, said Dr. Sharon Plon, a professor at Baylor College of Medicine in Texas. There are no large population studies that currently disclose genetic information to family members; only smaller, disease-specific ones do, she said.

Ellen Wright Clayton, a professor in Vanderbilt University’s law and medicine schools, said that people differ over privacy concerns and that researchers should not assume everyone is clamoring for genetic information.

But Scott Nelson, a patient advocate and 10-year survivor of pancreatic cancer, had this to say: “I’m going to want more information than you are going to want to give me.”

He noted that limiting disclosure to cases where the information is “actionable” is a problem because what’s not actionable today might be actionable tomorrow. And even if it’s not, he said, “the information may help me make decisions about how to live my life.”