Identical Twins Are Genetically Different

Posted by Pete Shanks April 5, 2011
Biopolitical Times

Twin studies have a solid place in scientific methodology for one simple reason: The DNA of identical twins is assumed to be identical, which removes an important variable from any experiment. Differences in their behavior, health and so on must therefore derive from the environment.

There have certainly been critics of this approach, and of the heritability estimates derived from it, but it has been a standard research tool for a century. New research, however, by a team at the University of Western Ontario led by Shiva Singh and Richard O'Reilly, found that the truth is more complicated. Monozygotic ("identical") twins do not have identical genomes.

"That assumption has been with us since the beginning of time," said Singh. "So this finding could be really revolutionary."

The researchers are investigating schizophrenia, which tends to run in families, though monozygotic twins are frequently "discordant for schizophrenia" — if one has it, about half the time the other does not. So they conducted a different kind of twin study, one that, instead of looking broadly at many twins, looked closely at a few.

They studied two families, each of which included one twin diagnosed schizophrenic, and used Affymetrix technology to evaluate both sets of twins and both sets of parents for about a million markers. Singh explained:

[A]ll of a sudden I realized … these guys have differences. They are not genetically identical.

Even more important, none of us have the same DNA we once did, and our DNA will keep right on changing:

Cells are dividing as we develop and differentiate. More important, these cells may lose or acquire additional DNA....The genome is not static. It's not that you got it from your parents and you will pass it along to the next generation, (and) in the meantime, they can determine everything you are going to be.

Or, as the abstract of the PLoS paper concludes:

The results argue for enduring genome-wide changes during cellular transmission, often ignored in most genetic analyses.

 

The implications for personalized genomic medicine could be profound. Possibly, a targeted therapy that worked last year might not work as well for the same patient later in life. But this is speculative, and it would be easy to overstate the implications of these findings. For instance, to say that "about 12 per cent of DNA can vary across individuals" sounds very dramatic, but actually monozygotic twins do tend to be rather alike (you may have noticed).

Singh and O'Reilly's team are moving on to whole-genome comparisons of their subjects, which should produce interesting results. Even there, care is needed: a recent report showed "1.1 million discrepancies" when the same human genome is sequenced with different technologies. That sounds terrible, but it's actually "very good congruence" over 3 billion bases. Let's just say that genomic research — and indeed the study of twins — has a long way yet to go.

Previously on Biopolitical Times: