By Michael Kahn
LONDON (Reuters) - Many genes linked to various cancers do not appear to raise the risk of getting cancer after all, according to an analysis of hundreds of studies published on Tuesday.
The findings highlight the need to exercise caution over the increasing number of studies associating common genetic variations with a range of diseases, said John Ioannidis of the University of Ioannina School of Medicine in Greece.
"The whole thing about genetic variations and links to diseases like cancer are very exciting, but the general public should be quite cautious about jumping to the conclusion that if they have a change in one gene or another they are doomed," Ioannidis, who led the study, said in a telephone interview.
"Genetic effects are very complex and very subtle and we need to know a lot more before we can make strong recommendations based on genetic profiles."
Ioannidis said his team had looked only at common genetic changes or polymorphisms, not at rare mutations, which in genes such as BRCA1 and BRCA2 significantly raise breast cancer risk. The rare form of these variants, for example, accounts for an estimated 5 to 10 percent of breast cancers.
Since early 2007, variations at more than 100 places on the genome have been linked to diabetes, heart disease and certain cancers.
The problem, researchers say, is that many of these genes typically interact in a complicated manner and their ultimate effects are influenced by the environment -- diet, exercise, smoking and other behavior -- in often poorly understood ways.
Ioannidis and his colleague Paolo Vineis of Imperial College London analyzed hundreds of published studies linking genetic changes to different cancers. They found that, out of 240 associations between a specific mutation and a cancer, only two genes involved in DNA repair and tied to lung cancer -- XRCC1 and ERCC2 -- turned out to be strong candidates for such a link.
"Most of the associations had weak or modest credibility," he said. That included PARP1 for breast cancer and CCND1 for head and neck tumors.
The problem is that on their own, the earlier studies fail to provide a complete picture and run the risk of drawing conclusions from too limited an amount of data, Ioannidis said.
This does not mean studies linking genes to cancer and the risk of other diseases have completely missed the mark, but rather that it takes a mountain of evidence to reach strong conclusions when it comes to the human genome, he added.
"Our study shows that it really takes a lot of research effort and many, many studies to be able to pinpoint a couple of associations," Ioannidis said.
(Reporting by Michael Kahn; editing by Maggie Fox and Tim Pearce)
