Why dogs are such different sizes

Dogs have a special feature among mammals. Their species is the one whose size is most malleable. Compare a Chihuahua and a St. Bernard! Naturally, geneticists wondered about the causes of such plasticity. Before the advent of genomics, however, they could only observe and put this fact into perspective. But access to the genome has unblocked the situation. Now, after more than twenty years of research, Elaine Ostrander of the Human Genome Research Institute, Bethesda, Maryland, and her colleagues have solved the mystery and then some…

The researcher investigated the question at the end of the 1990s, when genomics was still in its infancy and the geneticist relied on the comparison of short mitochondrial DNA sequences (1 DNA contained in mitochondria, the energy-producing factories in cells) . She therefore suggested that the dogs separated from the wolves on several occasions. Then, using the techniques of the time to study chromosomes, she highlighted hypervariable regions of the genome that might control the size of dogs. She then assembled a dynamic team in Bethesda to investigate these regions. So in 2007 his attention turned to a specific gene from one of these regions, IGF1. In fact, this gene encodes an insulin-like hormone that acts on the development of cartilage, which is involved in the growth of long bones. And its mutation, although very rare, has effects that are well known in medicine: growth retardation, intellectual disability and deafness. Did we find the solution? Elaine Ostrander didn’t seem convinced, because a few years later she and her team published a paper suggesting that size reduction in dogs actually depends on variants in at least six different genes. Everything seemed to get complicated.

Finally came the era of comparative genomics. The Bethesda team has started a large-scale work, manipulating up to 722 different genomes! In 2019 she counted around twenty genes that influence leg length, stature or size. And among them the gene IGF1 turned out to be the most important, accounting for more than 15% of the size variations.

A suspicious gene

So recently the team re-focused on this important gene. Important and weird! In fact, no one had ever managed to identify variants of a specific size… This time, the team focused on 13 species of canids, carefully selected using this criterion, from which they examined 1,431 genomes. And she didn’t just puzzle the gene IGF1, but also the adjacent chromosomal region, and this with fiendish precision. Again, the team did not discover any variants of the gene IGF1. But among ten point mutations listed in the noncoding sequences that delimit this gene in different breeds of dogs — chow chows, Afghan hounds, Tibetan mastiffs — the team noticed one that stands out: on the four “letters” — the bases A (adenine), T (thymine), G (guanine), C (cytosine) – distinguishing the various components of DNA (nucleotides), only two – T and C – are detected at this point. However, 75% of homozygous CC dogs (ie whose two copies of the relevant chromosome carry the C-base allele) have a body mass of less than 15 kg, while that of 75% of homozygous TT dogs is greater than 25 kg.

To make this finding as clear as possible, Elaine Ostrander’s team subtly chose a specific terrain, that of the Schnauzer. There are actually giant, medium and miniature schnauzers. Miniature Schnauzers are CC homozygotes while Giants are TT homozygotes. Why is the variation of a single nucleotide so important? The team then selected 51 dogs and compared their genotype, body mass and blood levels of the IGF1 hormone. The correlation is strong: small animals, mostly of the CC genotype, have very low IGF1 levels. This is a major advance because it explains the absence of variants in the gene IGF1. If a change in a non-coding sequence leads to a sharp decrease in the circulating hormone level, this is a change in the regulatory mechanism of the gene.

The sequencing of the part of the genome surrounding the mutation is therefore crucial. The latter is located in a sequence at the origin of an RNA that does not code for protein and happens to be complementary to the gene’s messenger RNA IGF1. This non-coding RNA can therefore bind to the messenger RNA and from there, depending on the sequence, inhibit the production of the hormone to a greater or lesser extent.

Great or minor ancestor?

However, one question remained. Which of the two alleles C and T – nicknamed “small” and “large” allele – is the ancestor, which is the derivative? To answer this, Elaine Ostrander’s team surveyed the bones of 33 dogs that lived between 11,000 and 100 years ago. Result: Nordic animals (at a latitude greater than 55° N) carry the large allele much more frequently than animals further south, mainly in the Mediterranean region. This is consistent with Bergmann’s rule, which states that warm-blooded animals tend to be larger in cold countries, where the surface-to-volume ratio is then more favorable to them. Thus, it was shown that the two alleles coexisted at that time. However, the researchers secretly thought that the large allele was an ancestor and the small allele a more recent derivative since the small breeds are the result of recent artificial selection during the last two centuries when there was a strong demand for small domestic animals in Europe and the United States.

Since the two alleles appeared to be of the same age in the dog, we had to go back in time. The Bethesda team first studied ancient and modern wolves and found that the small allele is present in fossils at low frequencies. It has been identified in the heterozygous genome of a 53,000-year-old Siberian wolf! It was therefore necessary to go even further back in time and turn to other canids – coyotes, red wolves, African golden wolves, black-backed jackals, side-striped jackals, dholes, American gray foxes, Magellanic foxes – whose weights vary from 5 to 35 kg. However, apart from the red wolves – the largest animals – all other canids were homozygous for the small allele. In other words, the latter is unexpectedly the ancestral allele.


The history of the alleles can then be reconstructed. In ancient small canids (more than 53,000 years ago), the small allele is ubiquitous. The large allele appeared at the same time as the lineage of wolves, large animals, appeared. Then around 40,000 years ago the domestication of the dog began, always with large animals. And recent selection of companion animals has resulted in an artificial preponderance of the ancestral small allele in companion dogs.

Leave a Comment