Dogs are Still not Wolves: Human Feeding Practices Have Shaped the Dog Genome

I have written many times about the argument, often made to support raw meat feeding or other alternative dietary practices, that dogs are basically carnivores and that because they were derived from wolf ancestors their optimal diet should be as much like that of wild wolves as possible. (see links to previous posts)*

On one level, this argument is an example of the Appeal to Nature Fallacy, which says that anything arbitrarily defined as “natural” must be benign or beneficial. In the case of diets for carnivores, this ignores the obvious facts that: carnivores in the wild don’t eat an optimal diet, they eat whatever they can catch or scavenge; malnutrition, parasites, broken teeth, and other harmful consequences of a “natural” diet are ubiquitous in wild carnivores; captive carnivores, protected from these and other hazards often live longer, healthier lives than their wild counterparts.

However, the “dogs are wolves and should eat like wolves” argument fails on another level, which is that dogs simply are not wolves. That should be obvious to anyone who tries to imagine a pack of pugs taking down and savaging an elk. But even on less obvious and dramatic criteria, the distinctions between dogs and wolves wrought by domestication. As I have discussed previously, changes in dentition, the GI tract, and the production of digestive specific enzymes illustrate the effects of artificial selection on the ability of domestic dogs to make use of a much wider range of food sources than wolves. A recent study has further explore the genetic changes behind increased production of amylase, an enzyme for digesting starch, that have accompanied initial domestication and subsequent selective breeding by humans.

Reiter T, Jagoda E, Capellini TD (2016) Dietary Variation and Evolution of Gene Copy Number among Dog Breeds. PLoS ONE 11(2): e0148899. doi:10.1371/journal.pone.0148899

As the authors explain,

Over time, and via cohabitation, the canine diet has been transformed from the carnivorous diet of its ancestor, the wolf, to a diet more closely matching that of omnivorous humans…This transformation increased variation in the domesticated dog’s diet, potentiating impact on numerous biological pathways.

Because humans populated a variety of habitats with different dietary staples, dog breeds from different places also consumed diets composed of unique combinations of food items. For many breeds, dietary changes resulted in increases in novel food constituents that may have required new, better, or more digestive mechanisms, thereby exerting differential selective forces on dogs living among different groups of humans. For example, starch digestion presented a new dietary challenge to which the dog likely adapted through alteration of three key genes in the starch digestion pathway

The study investigated the gene for amylase, and several other genes, and found that not only did the dog begin to produce significantly more amylase than the wolf after initial domestication, but that specific breeds of dogs produce different amounts based on the amount of starch in the diet of humans in the areas where those breeds were developed. In other words, changes in the nutritional needs of dogs have continued beyond initial domestication, showing the powerful impact of intensive selection, and these changes have been driven by what is eaten by humans particular breeds live with, reinforcing that until recently dogs have been shaped by human activity to adjust to our diet.

Consistent with studies with more limited datasets, we found that AMY2B CNV did vary with dietary starch intake. Dogs with high-starch diets…had a statistically significant higher mean CNV…compared to dogs with low-starch diets….

These findings expand upon the recent study of Axelsson and colleagues [ref in original], which found that AMY2B copy number is substantially increased in domestic dogs relative to wolves. This study presents evidence that in dog breeds that were exposed to starch-rich diets, positive selection continued to influence AMY2B copy number after this initial copy number expansion.

The information from this single paper does not, of course, tell us much about the optimal diet for dogs generally nor for any particular breed. It simply illustrates that human activity, including what we eat and, consequently, what our dogs eat, has shaped our dogs genetically and functionally as well as in the obvious anatomical ways. Dogs are not wolves, and their dietary needs are much more a function of what humans have historically fed them than of what carnivores in the wild eat.


* Raw Meat and Bones for Dogs: It’s Enough to Make You BARF

Repeat After Me: “Dogs are not Wolves”

One More Time: Dogs are not Wolves!

This entry was posted in Nutrition. Bookmark the permalink.

9 Responses to Dogs are Still not Wolves: Human Feeding Practices Have Shaped the Dog Genome

  1. kitty says:

    Would be more interesting if they looked at cats instead of dogs and compared cat genome with that of their ancestor Near Eastern wildcat or other small wildcat subspecies of Felis Sylvestris like European and Scottish wildcats.

    Not arguing the main point – the fact that wild diet isn’t necessarily optimal still makes sense, but as to the rest, the difference between dogs and wolves is probably a lot greater than between a domestic cat of no particular breed with a wildcat.

  2. skeptvet says:

    This would be interesting. I suspect there would be a lot fewer differences, of course, since selection hasn’t been nearly as intense on domesticated cats as on dogs. There is still, of course, little evidence on the optimal diet for cats, and extrapolations based on phylogeny are problematic even in the absence of intense artificial selection.

  3. Niall Taylor says:

    Brilliant post, as ever, Brennen, thanks. I wrote a piece on the RationalVetMed website about an article with similar findings a couple of years ago: Axelsson, E., et al (2013) The genomic signature of dog domestication reveals adaptation to a starch-rich diet Nature Vol. 495, pp. 360–364 ( Now I’ll be able to include this article too!

    There is way too much ‘heart’ and not nearly enough ‘head’ in the raw food debate.

  4. Vivian Ting says:

    Hi: Great article. If you were making food from your dog from scratch, what ratio of calories from carbohydrate, fat and protein would you recommend for a heathy senior dog that is not that active due to arthritis?
    Many thanks for your advice.

  5. skeptvet says:

    I would recommend getting expert advice about both macronutrient and micronutrient content. There are lots of boarded veterinary nutritionists who do such consultations. You can usually find one at the closest veterinary college, or online (e.g.,

  6. Mary says:

    Hello! Since dogs are omnivores, shouldn’t pet foods have a higher content of meat? I ask this because the majority of the ingredients in most pet foods are vegetarian, with only a low or very low proportion of meat (sometimes less than 5%…). So, they are basically vegetarian diets, rather than omnivore diets… Thanks!

  7. skeptvet says:

    Nutrition is just a lot more complicated than this. The proportion of each macronutrient (carbohydrate, protein, and fat) and the levels of various micronutrients as well as the particular sources used for these all make a difference, but it’s not correct to say “more meat=better.” In fact, I just saw a dog this week on a meat-based raw diet with dangerously low phsphorus levels, because meat contains too little phosophorus and the owner was not balancing the diet with other ingredients.

    If diets contain an appropriate amount of macro and micronutrients in an available form, the source (plant or animal) is less important.

  8. Mary says:

    Ok, I see, thanks! Do you have any complete recipe you can share with us for a healthy adult dog?? I would love to try a fresh diet! Thanks in advance!

  9. Pingback: Raw chicken/turkey bones - Forums - Yorkshire Terrier Community

Leave a Reply

Your email address will not be published. Required fields are marked *