Scientists claim they can predict whether someone will live to 100 with up to 85 per cent accuracy after developing a genetic test for healthy aging.

Researchers came up with the complex formula after studying the DNA of centenarians and identifying common genetic variations which may have contributed to their longevity.

They found a series of markers which appeared to have made people genetically predisposed to a long and healthy life, for example by helping to delay or prevent the onset of age-related diseases such as heart disease and dementia.

These genes could help explain why living beyond 100 appears to run in families, despite the importance of factors such as smoking and diet on our long-term health.

The test is between 60 and 85 per cent accurate, depending on the subject's age and is most reliable late on in life, meaning it is unlikely to be useful as an early predictor of an individual's lifespan.

Its ability to pick out long-lived groups who share certain genes could be useful to scientists hoping to find new treatments for age-related conditions.

Dr Thomas Perls, one of the lead authors, said: "Further study of these genetic characteristics may yield a better understanding of the genetic and biological bases of delaying or escaping age-related diseases and achieving longer survival."

The study, by Boston University researchers, expands on work in a 2010 paper which was published in the Science journal but later retracted due to problems with the data. In the new paper, which was independently validated by Yale University experts, an additional sample of participants with an average age of 107 was studied.

People who live beyond 100 are an ideal model for healthy aging because they generally do not begin to experience disability until their mid-nineties.

While genetic factors are only thought to influence our chance of living to 85 by about 20 to 30 per cent, the study suggests they could play a much greater role in the late eighties and beyond.

Researchers scanned the genomes of 801 centenarians and cross-referenced them with similar samples from 914 younger, healthy people to find DNA characteristics which were concentrated in the older group.

They identified 281 genetic variants which appeared to play a role in aging and found that they could be grouped into 26 different "genetic signatures", which characterized 90 per cent of participants.

Tim Spector, professor of genetic epidemiology at Kings College London, said: "This is an interesting study of rare centenarians - it suggests that genes are much more important in deciding if you live to 100 or beyond than if you live to 85. With hundreds of genes contributing, personal prediction is never going to be good.

"The big news is that there is not one gene for aging, it is literally hundreds of genes and some might be inherited in clusters but we do not yet know how they work.

"If we can work it out then there is the possibility of anti-aging therapies."