NEW YORK – The genomes of monozygotic twins differ on average by 5.2 early developmental mutations, and about 15 percent of them have a significant number of such mutations that are specific to only one of them, according to a new study.
In a paper published in Nature Genetics on Thursday, researchers at DeCode Genetics, the University of Iceland, and Reykjavik University said they observed instances where a twin was formed from a single cell lineage in the pre-twinning cell mass and also saw cases where a twin was formed from several cell lineages. Further, they said, CpG>TpG mutations increased in frequency with embryonic development, coinciding with an increase in DNA methylation, and the overall results indicated that allocations of cells during development shape genomic differences between identical twins.
Importantly, the researchers noted, phenotypic discordance between monozygotic twins has generally been attributed to the environment. This theory also assumes that the contribution of mutations that separate monozygotic twins is negligible. These data seem to contradict that assumption.
"One of the things that becomes important when you begin to look at things like this is the implications of these de novo mutations that separate identical twins," said co-corresponding author and DeCode Genetics CEO Kari Stefansson. "One of the implications is that the use of identical twins to separate genetics-versus-environment becomes somewhat disconcerting. You have to be careful when you assume that the differences between identical twins are simply because of environmental influences."
DNA from the blood of monozygotic twins was previously shown to have differences that may be due to somatic mutations, the authors noted. Such mutations are likely to be increasingly detectable as twins age, due to clonal hematopoiesis. To track when these mutations truly begin, the researchers looked at the earliest stages of human development and set out to put a timepoint on mutations that are specific to one twin that must have occurred after the initial formation of the zygote. At one to two weeks after blastocyst formation, a set of cells in the embryo are slated to become germ cells — a process called primordial germ cell specification (PGCS), they explained. Postzygotic mutations present in both the germ and somatic cells of twins most likely occurred before PGCS.
To estimate the number and timing of mutations differing between monozygotic twins, they searched for postzygotic mutations occurring in the somatic tissues of only one twin, looking at 381 pairs of discordant monozygotic twins. They timed these mutations by comparing whole-genome sequencing data from monozygotic twins, their offspring, spouses, and parents. In total, they found 23,653 postzygotic mutations that were specific to one twin, with a median of 14 postzygotic mutations differing between a pair of twins.
Using this data, the researchers were able to divide twin pairs into two groups, one where both twins were formed from the same cell lineages of the pre-twinning cell population and the other where they were not. This allowed them to determine the number of mutations that separate monozygotic twins, their type, and the timing of their occurrence. There was considerable variability in the number of postzygotic mutations. For example, 39 twin pairs differed by more than 100 mutations.
To find the pre-PGCS mutations, they also sequenced the genomes of offspring and spouses or partners of 181 monozygotic twin pairs to look for mutations present in the proband's offspring but not in their twin or spouse/partner. They found that 27,265 mutations absent from the proband's twin and spouse/partner were transmitted to their offspring. Of these, 582 mutations were found in the somatic tissue of the proband and were therefore most likely pre-PGCS mutations. On average, 1.3 pre-PGCS mutations were transmitted from a proband to each offspring.
Stefansson particularly noted that DeCode's whole-genome sequencing efforts for diagnostic purposes has shed light on how many devastating early childhood diseases are caused by de novo mutations. "What is more, these diseases are not just the most devastating syndromes of early childhood, they're also diseases like autism," Stefansson added. "So when you run into identical twins, one of them with autism and one without, up until now, people have always assumed the difference there is due to environmental influences that brought about autism. But that is a dangerous assumption. It could equally well be because of de novo mutations that happened just in one of the twins and not the other."
This doesn't mean that twin studies can't be useful for studying disease or determining the effects of environment versus genetics on the development of illness, though, he said. But it is important to be careful in the interpretation of the data. Stefansson also noted that the study demonstrates the importance of whole-genome sequencing in the study of disease, and its possible use as a diagnostic tool.