How much breed actually explains
The Darwin's Ark team surveyed the owners of 18,385 dogs and genotyped about 2,155 of them, then measured how much of each behavior could be attributed to breed. Here is the answer for all eight behavioral dimensions, strongest to weakest. Notice the ceiling. Breed never explains more than about a fifth of anything.
The aggression myth
Here is the number that should change how we talk about dogs. The behavioral factor closest to aggression, what the researchers call Agonistic Threshold, was explained by breed at about 9 percent. Its genetic heritability came in around 0.10.
Put plainly: the breed on a dog's paperwork barely predicts whether that dog will be aggressive. Two dogs of the same breed can sit at opposite ends of the scale. The thing that moves the needle is the individual animal and the life it has lived, not the label.
This is why breed-specific legislation keeps failing the evidence test. Banning a breed targets a category that explains less than a tenth of the behavior it claims to prevent. The honest tool is the one shelters and trainers already know works. Evaluate the dog in front of you.
Why breed feels more predictive than it is
Breed shapes how a dog looks far more than how it acts. We see a blocky head or a herding coat and our brain fills in a personality to match. Then expectation does the rest. We notice the Lab that loves water and the terrier that digs, and we quietly forget the ones that do neither.
The genetics back this up. The traits breed shaped most, trainability and toy drive, are exactly the ones dogs were bred to perform. The traits breed shaped least, friendliness and reactivity, are closer to universal dog nature. Selection wrote the working manual. It barely touched the temperament.
What this means for your dog
This is good news, and it is freeing. Your dog's future is not written in its breed. A well-socialized dog of a feared breed is a wonderful dog. A poorly handled dog of a beloved breed can be a difficult one. You have far more influence than the label does.
Use breed for what it is good for. It hints at energy level and the kind of work a dog might enjoy, and it flags the health conditions worth testing for. For who your dog is, watch the dog. Train them, socialize them early, and meet the animal in front of you on its own terms.
The same lesson is reshaping human genetics, where a single ancestry label was long mistaken for a verdict and turned out to be a weak one. Dogs are teaching us to read the individual, not the category. That is the whole point of building this atlas in the open.
Frequently asked questions
Does a dog's breed determine its personality?
No. In the largest study of dog behavior and DNA to date (Morrill 2022, 18,385 dogs), breed explained at most about 18 percent of any single behavioral trait, and for most traits far less. Breed is a weak hint, not a forecast. The individual dog, their training, and their environment account for the large majority of behavior.
Are some breeds genetically aggressive?
The evidence does not support it. In Morrill 2022, the behavioral factor most related to aggression (Agonistic Threshold) was explained by breed at only about 9 percent, with a genetic heritability around 0.10. That means a dog's breed tells you almost nothing about whether it will be aggressive. Aggression is overwhelmingly about the individual dog and how it was raised, socialized, and handled, not its breed label.
What behaviors does breed actually predict best?
Working and play traits, not temperament toward people. Breed had the strongest influence on trainability (biddability) and toy and fetch drive, each around 18 percent. These are the traits dogs were historically selected for. Breed had much weaker influence on sociability and aggression, which are closer to universal dog traits shaped mostly by the individual.
Should breed bans (BSL) exist if breed barely predicts aggression?
The genetics argue against breed-specific legislation. If breed explains roughly 9 percent of reactivity and has a heritability near 0.10, banning a breed targets a label that barely predicts the behavior it claims to prevent. Major veterinary and behavior organizations have reached the same conclusion from the behavioral side. The data points to evaluating the individual dog and its history, not its breed.
Why does breed feel so predictive if the science says it is weak?
Two reasons. First, breed shapes how dogs look far more than how they act, and we read a lot into appearance. Second, expectation is self-confirming: people notice the behaviors that match the stereotype and forget the ones that do not. The numbers cut through both. Across eight behavioral dimensions, breed never explained more than about a fifth of any of them.
Source: Morrill K, Hekman J, Li X, et al. Ancestry-inclusive dog genomics challenges popular breed stereotypes. Science 2022;376(6592):eabk0639. doi:10.1126/science.abk0639. Behavioral data from the Darwin's Ark project (CC0). Per-breed temperament profiles appear on each breed page.