nelson provaziThis study challenges what is commonly taught in high school classrooms: “Characteristics such as skin color depend on the complex relations among many genes, which makes it practically unfeasible to identify the phenotype of an individual on the basis of his genetic makeup (genotype).” Although the first statement is correct, Brazilian researchers have shown that it is possible to establish the skin’s pigmentation on the basis of genes. So far, the researchers have been 60% correct.
This finding, made by the group led by geneticist Maria Cátira Bortolini from the Federal University of Rio Grande do Sul (UFRGS) would have been impossible if we were not on the brink of the personalized gene era, thanks to the fact that it has become financially feasible to sequence a full set of genes of individuals and to make this information available on web databases. Thanks to this availability, the Brazilian team did not have to sequence anybody’s DNA. The team did its research work with genome data found in public databases around the world.
The team from UFRGS used information from 30 individual genomes. Some of the genomes were of well-known people, such as American geneticists Craig Venter and James Watson, which allowed the team to confront the genetic data with the phenotypical characteristics (appearance). Other genomes came from anonymous individuals, whose phenotypes were estimated on the basis of physical characteristics of the ethnic groups to which these individuals belonged. The researchers also analyzed the genomes of a paleo-Eskimo and four archaic hominids: three Neanderthals and one Denisovan hominid – actually, the latter was a woman that lived in Siberia forty thousand years ago, possibly belonging to a previously unknown Homo species.
To collect useful data for the research from this enormous alphabet soup – each genome is comprised of two sets of 23 chromosomes totaling 3 billion pairs of A, T, C and G nitrogenous bases – the researchers resorted to a method developed by geneticist Caio Cesar Silva de Cerqueira, the first author of the study, accepted for publication in the American Journal of Human Biology. “This work is the sequence to my doctorate thesis on skin color genes in human populations,” says Cerqueira, whose advisor is Cátira.
The magnitude of this task cannot be underestimated. “One of the biggest difficulties was finding a way to analyze this huge volume of data at the same time,” says Cerqueira. “As far as we know, there is no statistical apparatus that can do this in a simple manner.”
This is why the team’s first mission was to focus the analyses on stretches of DNA that could provide the estimates with greater reliability. The group basically worked with genetic differences referred to as single nucleotide polymorphisms or SNPs. The SNPs represent genetic differences in which only one letter in the sequence was changed. “We had to filter the data to work with only more tangible and real data,” explains Cerqueira.
Data selection
The starting point was to identify 346 SNPs distributed over 67 genes, pieces of inactivated genes (pseudo genes) and intergene regions (not all DNA segments constitute genes; some only take up space in the sequence and their function is not clear yet). All these SNPs were in regions of the genome associated with hair, eyes, and skin pigmentation. The next step was to check which SNPs were described, in terms of their genetic effects, in the available literature. Only 124 of the 346 SNPs were thus described.
There was another problem: the genome is comprised of two copies of each gene, one from the father and one from the mother. It is very complicated to predict the effect that the combination will have on the organism when the versions of the gene differ between themselves. This is why the researchers concentrated on the SNPs whose alleles (alternative forms found simultaneously in the organism) were found in two copies of the same stretch of each genome. “We lost a lot of information when we did this, but we decided to take this more conservative approach,” says Cerqueira.
The methodology that the researchers employed is the appropriate one to establish whether the individual has freckles or not, a pigment accumulation common to blonds and redheads. The success rate in the prediction for the 11 genomes whose phenotype was well known (though the identity of the owner was known) corresponded to an impressive 91%.
Nelson ProvaziHowever, as the subtleties increased, the success rate dropped. The methodology predicted the skin color correctly in 64% of the cases – the skin was divided into two categories: light and dark. The prediction rate corresponded to 44% for hair color (black, brown, red, and blond), and 36% for eye color (black, brown, green, blue)). When all the characteristics were taken into account, the average success rate corresponded to 59%.
The researchers also increased the success level by including 19 genomes of individuals whose ethnic background allowed the probable phenotype to be estimated. The success rate changed somewhat with a base of 30 genomes. The levels dropped slightly for freckles (83%), skin (60%) and hair (42%). They increased for eyes (67%), thus raising the final average to 63%.
The first conclusion from the study is that it is no longer possible to state that predicting physical features based on an analysis of the DNA is impossible. The second conclusion is that there is still a lot of ground to be covered before the level of accuracy is improved to the point of making this test useful.
This technology could cause a revolution in the forensic sciences, for example. Based on a DNA sample found at the site of the crime, the police could create a detailed picture of a suspect. We are still far from this technological stage but Cátira says that the “what if we could?” stage has already come to an end and we are reaching the “ how are we going to do this?” stage.
“The biggest challenge is understanding how the interaction between genes and their alleles and proteins works,” says Cátira. “In other words, we have to understand to what extent the effect of an allele found at a specific point on the route is altered by the existence of another variant in another gene of the pigmentation network. Studies about these connections are just beginning and we have no idea how everything is connected to result in a given phenotype,” says the researcher from the UFRGS.
An additional complication is that the epigenetic effects have to be taken into account – epigenetic effects are the influence of environmental factors on the patterns of expression of certain genes without altering the DNA itself. “The challenges are huge,” says Cátira. “However, scientific knowledge is growing exponentially and so I have hopes that major advances will occur in the upcoming years.”
Portraits of evolution
Even though technology has not yet reached a point where it can be useful to police investigations, researchers have already begun to use it to gain a better understanding of the evolution of the Homo genus. Studies such as this one help verify to what extent the differences in pigmentation among human groups is a result of the pressure exerted by natural selection or whether it consists of variations that appear at random and are therefore neutral from the evolutionary point of view.
In a previous study, linked to another characteristic, Cátira’s group showed that a gene associated with the configuration of limbs in human beings remains exactly the same in more than one hundred DNA samples coming from people all around the world. This gene accrued 16 alterations since human beings and chimpanzees separated on the evolutionary tree but remained identical in the Neanderthals – Homo neanderthalensis, a species that is kin to the Homo sapiens, with whom it shared an existence until 30 thousand years ago, prior to disappearing. The conclusion is that this gene is extremely important and this is why it remained the same for such a long time, in spite of the evolutionary pressure exerted on it.
Nelson ProvaziResearchers can now conduct similar analyses in relation to skin, eye and hair pigmentation to see what kind of evolutionary role the genes related to these characteristics might have played. Before any DNA analyses had been conducted, many researchers had already thought that there must have been significant evolutionary pressure on humans living under the scorching African sun to have more melanin in the skin – and therefore more protection against harmful solar radiation – while the humans that lived in the North of Europe would rarely need large amounts of this pigment to protect them from damage due to sun exposure. Studies such as those conducted by the group headed by Cátira help us understand what shaped these and other adaptations.
One of the astonishing results of this new research study showed that there might have been differences in skin and hair color among the Neanderthals. The analysis of genetic variants of the Neanderthals – sections of the genome of three females were overlapped to obtain the full genome – suggests that one of the females was a redhead and two had brown hair and darker skin. All three women had brown eyes.
This study is in contrast with a previous study conducted by Carles Lalueza-Fox, of Barcelona’s Pompeu Fabra University. In a study published in 2007 in Science, the Spanish group showed that the genetic material of two Neanderthals – one from Spain and the other from Italy – had genetic alterations similar to the ones that determine light skin and red hair in human beings. “We know only a few genomes or sections of genomes of these hominids and, even so, this variation shows up,” says Cátira.
If the analysis conducted by the team from UFRGS is correct, it indicates that pigmentation among the Neanderthals might have varied in a similar manner to that occuring in human beings. “This would be quite reasonable and would indicate that this characteristic might be typical of the Homo genus and not of the human species,” says Cátira. She warns, however, that one must be careful when interpreting this data. “We cannot disregard methodological problems, such as contamination with human DNA and the exchange of post-mortem bases when genomes of extinct species are sequenced,” she adds.
This comment touches upon an important point: there are limitations in the analysis of the genetic material of fossils. For example, it might never be possible to investigate the DNA of the first human beings who settled in what is now Brazil and who possibly crossed over the Behring Strait from Asia between 20 thousand and 12 thousand years ago. “The problem is that the climate here does not preserve the DNA in the same way as DNA is preserved in the colder climate in Europe,” explains Fabricio Rodrigues dos Santos, a biologist from the Federal University of Minas Gerais. “If the researchers are lucky and find a well-preserved skeleton at some special site in South America that is more than 8 thousand years old, and has the DNA, then it might be possible to predict some phenotypes.”
As for Luzia, the human fossil found in the 1970s by French archaeologist Annette Laming-Emperaire in Lagoa Santa, State of Minas Gerais, this is the oldest human fossil found in the Americas; archaeologists estimate that it is between 11,400 to 16,400 years old. “In the case of Luzia, I would say that analyzing her DNA is impossible; various researchers have already tried; they sent material to the United States and Europe, but were never able to generate any DNA sequencing,” says Santos. Some gaps will inevitably remain, no matter how powerful genetics may be in terms of shedding light on the past of human beings, at least until the next scientific revolution.
Scientific article
CERQUEIRA, C.C.S.; et al. Predicting Homo pigmentation phenotype through genomic data: From Neanderthal to James Watson. American Journal of Human Biology. In press.
