SNPs as Windows on Evolution

Single nucleotide polymorphisms--variants in DNA sequences better known as SNPs and pronounced snips--provide a shortcut to comparing genes and genomes within and among species. The need to study SNPs has spawned a number of companies aimed at matching SNP patterns to disease risks. A few other organizations, however, are taking a broader view: mining SNPs for clues to human diversity and evolution.

Association studies that correlate SNP patterns to disease risks are straightforward. Clues to the past can be subtler, but they are found at all levels of evolution--from the great branching points of speciation, to ones among the primates, to those within modern human populations. And a general consensus has emerged: Humans haven't changed much since coming out of Africa long ago.

"This is a tremendously exciting time to study human variation, because we finally have enough tools and infrastructure to correlate genotype to phenotype. The human genome sequence is the foundation to investigate human sequence variation," says Eric Lander, director of the Whitehead Institute Center for Genome Research in Cambridge, Mass. And that variation reflects evolutionary processes that molded the modern genome. Lander was one of several speakers on these aspects of SNPs at the Fourth International Meeting on Single Nucleotide Polymorphisms and Complex Genome Analysis held in October at the Wenner-Gren Foundation in Stockholm, Sweden.

Single bases in a DNA sequence that differ in at least 1% of a population, SNPs pepper the human genome unevenly, accounting for about one in every 1,250 bases. Sets of certain SNPs are often inherited together more often than would be predicted by their individual frequencies, a phenomenon called linkage disequilibrium that divides the genome into blocks of information called haplotypes.1

Nonsynonymous SNPs Mark Speciation

Natural selection has left an imprint on SNP distribution. Most are in regions of the genome that do not encode protein, including the vast deserts that separate genes, and the introns within them, according to SNP Consortium Ltd. data provided by Lincoln Stein, associate professor of bioinformatics at the Cold Spring Harbor Laboratory in New York, where SNP consortium data is stored (snp.cshl.org).

Of the SNPs in protein-encoding genes, few change the protein. "More than 99% of human SNPs are not associated with biology. Only 2,000 SNPs change an amino acid sequence or a regulatory site. That is a few thousand out of 2-3 million SNPs cataloged so far," said J. Craig Venter, founder of Celera Genomics Group in Rockville, Md., in a keynote speech at the conference. The paucity of variation in protein-encoding genes compared to elsewhere in the genome makes evolutionary sense, for if a SNP occurs in a region that does not affect the phenotype, then it should be unaffected by natural selection and remain. Conversely, when a SNP is retained in a protein-encoding gene over time, it may be because the variant somehow helps the individual successfully reproduce.2

SNPs within protein-encoding genes are of two types. The degeneracy of the genetic code means that some SNPs are synonymous, not changing the encoded amino acid. A change of a CCA mRNA codon to CCG, for example, does not alter the designation of the amino acid proline. "Synonymous mutations occur only with time and are not under selective pressure," explained David Liberles, an assistant professor at the bioinformatics center at Stockholm University. Such a change is also known as a silent mutation. Nonsynonymous SNPs, in contrast, do alter the amino acid and therefore might be subject to natural selection. Those nonsynonymous SNPs that survive the weeding out of natural selection, and may actually have been selected for, are the foundation of the Adaptive Evolution Database (TAED).3 The idea behind TAED is that nonsynonymous SNPs that natural selection has retained mark key events in evolution.

To create TAED, Liberles and his team scanned GenBank www.ncbi.nlm.nih.gov/Genbank for genes encoding 26,843 protein families found in more than one species, indicating evolutionary antiquity. They then identified genes with the telltale nonsynonymous SNPs to generate "a raw list of potentially adaptively evolving genes." The researchers used a ratio of nonsynonymous DNA base changes to synonymous changes. A gene that has a high ratio indicates positive natural selection. The next step was to see which protein-encoding genes have high ratios in species that lie at the nodes of evolutionary trees, where an important change fueled branching of the tree. The coexistence of a SNP ratio indicating positive selection with a branchpoint signals evolution caught in the act, when a new protein variant that offered an advantage became fixed in the gene pool.

With a dose of imagination, considering the nature of the protein against the backdrop of an evolutionary tree, SNP ratios can reveal possible evolutionary scenarios. "The ratios are ... useful starting points for generating stories about the interactions between protein sequences and the Darwinian processes that shape these sequences. These stories help us understand how these sequences contribute to the fitness of the host," said Liberles. For example, nonsynonymous SNPs in the gene that encodes a plasminogen activator protein in the saliva of vampire bats may signal evolution of the ability to thin the blood of prey. The divergence of vertebrates on the animal family tree correlates to a high ratio for several genes that encode proteins involved in immunity. The genes that encode the satiety factor leptin, the double-muscle protein myostatin, and cellular adhesion molecules also undergo change at evolutionary crossroads. "At a biological level, the dataset generated here can be mined to provide global pictures of how evolution has occurred," Liberles summed up.

Considering Primates

SNPs reveal that Homo sapiens is a young species, with so little time having elapsed since origin that humans are what Venter calls "virtual identical twins." "The amount of variation in the human population is way less than expected for a population of 6 billion," concurred Andrew Clark, a professor of biology at Pennsylvania State University, indicating a rapid expansion.

"The poverty of variation tells us about the structure of the early human population. The rate of heterozygosity per nucleotide position is closely related to the size of a population and the mutation rate. Humans are a small population that grew large fast, from 10,000 founders in Africa 3,000 generations ago. Most of today's variation is the very same variation we walked out of Africa with. We are an extremely closely related species," said Lander.

Comparison to the closest relatives highlights humans' uniformity, reported Svante Paabo, director of the department of genetics at the Max-Planck Institute for Evolutionary Anthropology in Leipzig, Germany. "We looked at 10,000 base pairs in a nontranscribed region among 70 individuals, including 30 chimps from West, Central, and East Africa, and determined the mean number of differences. There were three to four times the number of differences between two random chimps than two random humans," he said. And despite the fact that one gorilla or orangutan might look much like any other to humans, they too are more variable.4

Chimps and humans may be very similar to each other at the genome sequence level, but the differences emerge in gene expression. Paabo's group considered relative levels of expression of 20,000 genes in humans, chimps, and macaques. "The results suggest that whereas the rate of change of the transcriptome in the liver and blood has been similar between chimpanzees and humans, the rate of overall transcriptome change in the human brain has accelerated about threefold. A number of genes that differ significantly in their expression levels between humans and chimpanzees have been identified," Paabo said. Other distinctions between humans and chimps might provide important health clues. Chimps do not suffer from malaria, bronchial asthma, rheumatoid arthritis, or AIDS. Epithelial cancers, such as of the breast, colon, and lung, are common among humans but extremely rare in chimps. Perhaps clues to the etiology of these conditions in humans lie in gene expression differences between humans and primate relatives.

Clues Among Modern Humans

The theme of similarity among humans was brought to modern times by Kenneth Kidd, a professor of genetics at Yale University. "Much of our variation occurs within all human populations. This suggests that a reduction of genetic variation associated with speciation may have occurred. In this way, we are just a very weird ape," he said. Kidd heads the Allele FREquency Database--ALFRED--which concentrates on those variations that do account for the obvious differences between a Swede and a Nigerian, a native American and an Asian.5 ALFRED, described on its Web site as a "work in progress," nonetheless considers 180 sites in the genome that are present in two alternate forms in different frequencies in at least six populations. Focusing on 14 of those genes that vary in their degree of linkage disequilibrium in different populations paints a clear portrait of an origin in and emigration from Africa. Said Kidd, "Homo sapiens clearly evolved in Africa. By 100,000 years ago, there was lots of genetic variability, and a reasonably dispersed geography. The only way to walk out was through northeast Africa. By 40,000 years ago, it was that group that got around the world." Since then, Founder effects have sequestered certain gene combinations as people tended to stick to their own kind, coupled with random genetic drift that sampled and amplified certain haplotypes. "In most cases, the same few haplotypes are the most common in the population--but not in Africa. Almost all of the haplotypes are due to Founder effects in coming out of Africa from what was a random gene pool," he explained.

Research at Genaissance Pharmaceuticals Inc. in New Haven, Conn., also points to an exit from Africa followed by rapid expansion. Although the focus of this company is to identify haplotypes (SNP combinations) that predict response to specific drugs, the approach also trips over signposts of human evolution.6 A recent paper reported on 313 genes in 82 people and a chimp as a control, revealing 4,304 haplotypes.7 The researchers surveyed African Americans, Asians, European Americans, and Hispanic-Latinos. At the Stockholm meeting, J. Claiborne Stephens, executive director of population genomics at the company, upped the numbers to 724, with 11,209 SNPs analyzed, and a gorilla added to the mix (The ape contingent allows identification of the oldest human alleles).

Again, the genome points clearly toward an African origin. "African Americans have the largest number of population-specific SNPs. Overall 56% of the SNPs are population-specific, and 21% are 'cosmopolitan,' which means that both variants are seen in all four populations," Stephens said. For many of the genes studied, one haplotype predominated--and the simplest explanation for that, Stephens concluded, is a recent and rapid expansion of the human gene pool.

1. L. Pray. "The promise that haplotypes hold," The Scientist, 15[23]:21, Nov. 26, 2001.

4. S. Paabo. "The human genome and our view of ourselves," Science, 291:1219-20, 2001.

5. M.V. Osier et al., "ALFRED: An allele frequency database for diverse populations and DNA polymorphisms--an update," Nucleic Acids Research, 29:317-9, 2001, www.info.med.yale.edu/genetics/kkidd or alfred.med.yale.edu/alfred/index.asp.

6. J. Wallace. "Personalized prescribing," The Scientist, 15[12]:10, June 11, 2001.

7. J.C. Stephens et al., "Haplotype variation and linkage disequilibrium in 313 human genes," Science, 293:489-93, July 20, 2001 and J.C. Stephens (Correction), Science, 293:1048, Aug. 20, 2001.

This is an interesting article, since it illuminates how evolution leaves its traces in the history of life. Note that the author of this report feels no need to defend the fact that evolution has occurred.

Of course, it's always possible that creature after creature was created, built from scratch, with gene sequences nearly showing what looks like random variations from creatures that look like ancestors.

Arguing against a rich history of descent with modification (aka evolution) in the face of this is like looking at someone standing in fresh snow, with a line of footprints behind him, and insisting that he didn't walk to his present location, but was dropped in place from the sky, and all those footprints are pure coincidence.

Amazing. No thought that we may be "extremely closely related" and that the variation hasn't changed since the human gene pool since "we walked out of Africa" because evolution didn't actually happen.

Most are in regions of the genome that do not encode protein, including the vast deserts that separate genes, and the introns within them, ...

For you biologists out there, is there some reason that biologists consider non-protein encoding sequences as having no function? Is it possible that these sequences have a function which we have not yet discovered? To use a computational analogy, there are varying layers of code each built upon each other, BIOS/OS/GUI/Applications. Is it not possible that DNA is the same? Protein encoding may be just one of many layers of code stored in DNA.

(Finally looking in after a few days...)

"The poverty of variation tells us about the structure of the early human population. The rate of heterozygosity per nucleotide position is closely related to the size of a population and the mutation rate. Humans are a small population that grew large fast, from 10,000 founders in Africa 3,000 generations ago. Most of today's variation is the very same variation we walked out of Africa with. We are an extremely closely related species," said Lander.

Amazing. No thought that we may be "extremely closely related" and that the variation hasn't changed since the human gene pool since "we walked out of Africa" because evolution didn't actually happen.

Except that there has been variation (read the article again -- carefully). What is found is that the most variation occurs among humans in Africa, and lesser amounts in groups that are thought to have migrated from Africa into Europe, Asia and the Americas. What the researchers find is 100,000 years worth of accumulated changes in populations that are still in Africa, but 40,000 years or fewer worth in populations outside Africa.

Why the difference? Why not all different amounts of variation in different populations?

I'm not sure whether I fall in that category, but I see nothing in this article that proves we're all descended from a clump of amino acids. In fact, one could argue that the very lack of diversity in the human genome actually supports the idea of special creation.

Good point. The problem is, how would this notion account for the fact that humans have, in common with apes, monkeys, and in fact all primates, identical copies of a broken gene for the production of vitamin C, all of which are broken in precisely the same way?

The evolutionary account is that all the primates inherited the gene from a common ancestor, which didn't need vitamin C while living on fruits and leaves. The creationist account can only be that it pleased a Creator to make all primates with identically-mangled versions of a gene that is functional in so many other creatures, for reasons that no one can explain.

It's possible, but...

It's kind of ironic that creationists (I'm not counting you in that group) demand lots of air-tight evidence before admitting to the existence of any sort of biological change, but are willing to spin all kinds of tales of possible functionality with no evidence to support them in any way.

There's lots of evidence that translation from DNA to RNA to proteins occurs. There's little, if any, evidence that any of the DNA not translated into RNA or proteins has any effect on the cell at all. And indeed, the DNA that's thought to be unused turns out to change at a certain background rate, corresponding to the frequency of random substitutions. DNA that has a useful function doesn't change at that rate, because there will always be some changes that have an adverse effect.

Indeed, measurements of the rate of change in DNA and other molecules has been cited as evidence that they do have some function. I recall an article citing the fact that proteins in the eyes of cave-dwelling fish changed at a rate slower than the background rate. This was cited as evidence that those proteins, although not used for vision, had to have some use.

Decades ago, before the mechanisms of heredity had been elucidated, scientists thought only proteins were complex enough to store hereditary information. The DNA was thought to be very simple, assumed to be an unending series of ATGC or something like that, and serving only to balance the alkaline proteins with which it was associated. After the central nature of DNA was discovered, it was decided that the proteins were there to balance the acidity of DNA.

It's possible that there's still stuff to be discovered. Indeed, the job of the research scientist depends on this fact. However, discoveries depend on evidence.

Got any?

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.