Sniffing out SNPs
First, scientists looked at the DNA from a bunch of different dogs. But they didn't look at every single A, C, T, and G (nucleotide) in the DNA. That would have cost a fortune. Instead, they focused on some special spots.
The vast majority of DNA in one dog is the same as the DNA in another. Fido's DNA looks pretty much the same as Lady's.
But there are some differences, and they are important. For example, Fido may have CATTA at a certain position, whereas Lady may have CAGTA. These differences are called Single Nucleotide Polymorphisms (SNPs).
Some of these SNPs affect something in Lady's or Fido's body while others do not. Whether they have an effect or not, scientists can use these SNPs as clues to track which parent a chunk of DNA came from.
Dogs (and people) have two copies of DNA, one from mom and one from dad. Let's say Fido and Lady have a puppy called Duke, and scientists look at Duke's DNA. If Duke has CATTA and CAGTA, scientists can tell where each copy of DNA came from. CATTA came from Fido and CAGTA from Lady. SNPs allow scientists to track pieces of DNA.
That's useful, but the scientists that figured out short-leggedness were not looking at dogs that were related to each other. They were looking at many dogs from many different family lines. So, how did they use the SNPs?
When animals in the same species share a trait (like leg length), there are often a bunch of different SNPs associated with those traits. These SNPs are usually clustered close together.
So the scientists figured that short-legged dogs all probably have the same letter at a particular SNP. And that letter is probably different in long-legged dogs.
The scientists looked at a lot of different SNPs in a lot of different dogs. At the end of the day, they could say something like, "All short-legged dogs have an 'A' at SNP-52, but long-legged dogs have 'G'."
Chasing down the right gene
That was great information but it wasn't necessarily that informative on its own. SNPs don't always tell scientists what part of the DNA is actually causing something like short legs. SNPs just told scientists that whatever was causing it was close by.
Understanding this requires an understanding of a concept called "recombination." Some people consider recombination a shuffling of DNA.
Where DNA is swapped in
recombination is random.
Think of getting DNA from parents like this. Everybody receives a book from mom and an almost identical copy of the same book from dad. As sperm or eggs are made, "recombination" occurs, which is like ripping out the last 100 pages of each book and attaching it to the other. The whole book is still there, but some pages came from mom's book and the rest from the dad's.
Here's the important thing about recombination: pages that are close together will probably stay close together as they are passed through the generations. This is because where recombination happens is totally random. It might happen on page 16. Or it might happen on page 973, 242.
So Lady's DNA from page 52 and page 53 will probably be inherited together by all of her ancestors. But something on page 1,000,000 will have less of a chance of being inherited with page 52.
That's because it's much less likely for recombination to occur between pages 52 and 53 than 52 and 1,000,000. To separate pages 52 and 53, ripping could only occur right between pages 52 and 53. But to separate pages 52 and 1,000,000, ripping could occur at page 54 or page 55 or page 56 . . . and so on to 1,000,000.
So, DNA sequences that are close together (like those on pages 52 and 53) are likely to be inherited together. If the DNA that is causing short legs is on page 53, odds are that it was inherited along with page 52 by all short-legged dogs.
To summarize, scientists found that all short-legged dogs had a certain SNP. This told them that the gene causing short legs was nearby because the SNP and the gene were probably inherited together. While the SNP narrowed down where to look, it didn't tell them exactly what was causing short legs.
But once scientists were close, they started looking at all the nucleotides in the DNA nearby to see if any changes in DNA made sense. For instance, perhaps a DNA change caused a protein to change, and that may have been the cause.
In the case of the short-legged dogs, it was much more interesting. They found an insertion of FGF4 that was present in breeds with short legs but not breeds with long legs.
This means that FGF4 was copied and inserted back into the dog DNA in a different spot. It's like taking page 795, where the original FGF4 was, copying it, and then sticking the copy into the book at page 53. From then on, FGF4 would almost always be inherited with the SNP on page 52.
There it is. Geneticists used SNPs and the principle of recombination to hunt down FGF4.
Figuring out that an extra copy of FGF4 causes short legs in dogs wasn't easy. Scientists had to use all their tools to find it.