In recent years, biology has been thrown around like a football. Activists in numerous areas invoke it - those against food science say scientists are tinkerers with no expertise while those against pesticides claim that the biology clearly shows what your grandparents ate made you obese.

Like in most scientific hot button topics, avoid the fringes. Epigenetics does what it does, but it is not determining your weight. For all the confusion, you have to blame epidemiologists who know just enough science to be wrong. If a pregnant mother is undernourished, epidemiologists have found that the baby has a higher risk of being obese. Then they blame epigenetics without being able to say what those are. Its genetic "memory", which unfortunately veers into water memory and homeopathy levels of evidence.

Regardless of the evidence, a new paper claims this 'memory' of nutrition during pregnancy can be passed through sperm of male offspring to the next generation, increasing risk of disease for grandchildren as well. Yes, 'you are what your grandmother ate'.

Where does it end? Why not great-grandparents? What if the people who got off that boat were fat? It's a shocking abuse of inheritance evidence. We inherit half of our genes from our mother and half from our father, but epigenetic effects, whereby a 'memory' of the parent's environment is passed down through the generations, are so little understood that papers claiming they can attribute Chemical X to Effect Y are doing a shocking amount of damage to public acceptance of science. The only understood epigenetic effects are caused by a mechanism known as 'methylation' in which the molecule methyl attaches itself to our DNA and acts to switch genes on or off.

A paper in Science notes that environmentally-induced methylation changes occur only in certain regions of our genome and that these methylation patterns are not passed on indefinitely.

Researchers led by the University of Cambridge and Joslin Diabetes Center/Harvard Medical School, Boston, used mice to model the impact that under-nutrition during pregnancy had on the offspring and to look for the mechanisms by which this effect was passed down through the generations. The male offspring of an undernourished mother were, as expected, smaller than average and, if fed a normal diet, went on to develop diabetes. Strikingly, the offspring of these were also born small and developed diabetes as adults, despite their own mothers never being undernourished.

Professor Anne Ferguson-Smith, from the Department of Genetics at the University of Cambridge, says, "When food is scarce, children may be born 'pre-programmed' to cope with undernourishment. In the event of a sudden abundance in food, their bodies cannot cope and they can develop metabolic diseases such as diabetes. We need to understand how these adaptations between generations occur since these may help us understand the record levels of obesity and type 2 diabetes in our society today."  

To see how this effect might be passed on, the researchers analysed the sperm of offspring before the onset of diabetes to look at the methylation patterns. They found that the mouse's DNA was less methylated in 111 regions relative to a control sperm. These regions tended to be clustered in the non-coding regions of DNA – the areas of DNA responsible for regulating the mouse's genes. They also showed that in the grandchildren, the genes next to these methylated regions were not functioning correctly – the offspring had inherited a 'memory' of its grandmother's under-nutrition.

When the researchers looked at the grandchild's DNA, they found that the methylation changes had disappeared: the memory of the grandmother's under-nutrition had been erased from the DNA – or at least, was no longer being transmitted via methylation.

"This was a big surprise: dogma suggested that these methylation patterns might persist down the generations," adds co-author Dr Mary-Elizabeth Patti from the Joslin Diabetes Center and Harvard Medical School, Boston. "From an evolutionary point of view, however, it makes sense. Our environment changes and we can move from famine to feast, so our bodies need to be able to adapt. Epigenetic changes may in fact wear off. This could give us some optimism that any epigenetic influence on our society's obesity and diabetes problem might also be limited and/or reversible."

The researchers are now looking at whether epigenetic effects no longer have an impact on great-grandchildren and their subsequent offspring. So, if it's true that 'you are what your grandmother ate', it might not be true that 'you are what your great grandmother ate' - or even people on the Mayflower.