You might be surprised at how important this stuff is in the biosciences. It's used as a "blocking agent" in Southern blotting, a common method for identifying a DNA of interest in a forest of DNA molecules, and as a "carrier" when trying to introduce DNA into yeast. DNA tends to interact with DNA; in cases where you want two kinds of DNA to interact in a general, not entirely specific, manner, you want DNA that comes from a species that isn't closely related to the organism you're working with. If you're working with human DNA, salmon sperm DNA will fit the bill. If you're working with salmon, of course, you'll have to find another source of DNA. In European genetics labs, you'd be more likely to find a bottle of herring sperm in the freezer.
You also want a relatively cheap source of DNA. Male salmon release their load into the water, so they need to make a lot of the stuff. As is the case throughout the animal kingdom, sperm has a high concentration of DNA (i.e. the cell doesn't contain much more than a nucleus), so it's a good source of DNA.
The dry flakes are difficult to dissolve in water, so you've got to boil the stuff. I had to prepare a few milliliters last weekend. Oddly, after all the processing and purification it took Sigma Labs to put 1 gram in a plastic bottle (at about $100 a pop), the distinctive aroma of salmon permeated our lab upon boiling.
Interestingly, if you examine the history of DNA science, salmon sperm has played an important role from the very beginning. Crick and Watson figured out the structure of DNA, but a dude by the name of Friedrich Miescher worked out its basic chemistry nearly a century prior. You can read the story in detail. Miescher, apparently, did fine and meticulous work, and understood that this substance ("nuclein", back then) must be of importance in the cell.
Salmon sperm (from the Rhine) provided a good source of DNA for Miescher's studies. Back then, if you wanted to characterize some chemical, you'd heat it and boil it and torture it, trying to register a weight change in, say, a phosphorus-absorbing substance. That way, Miescher worked out that DNA is 3% phosphorus. Needless to say, that approach required copious quantities of starting material.
Since sperm is mostly nuclei (which is where the DNA resides), it made sense that it was involved in fertilization. Given the techniques of the day, though, it seemed that nuclein from salmon was no different than nuclein derived from any other organism, so Miescher looked elsewhere for the genetic essence, even speculating that the chiral nature of protein might hold the key (40 chiral centers in a protein mean 1,099,511,627,776 arrangements!).
Did you know there's actually a disease, trimethylaminuria, where the patient gives off a fishy odor? According to Wikipedia, "Living with TMAU can be challenging, and TMAU can adversely affect the livelihood of the people who have it, as well as their families." Yikes!