A comment from a reader on a recent zebrafish-related news post:
IMO it's a waste of money and scientists. We should only focus on mammals, because humans are no fish. It won't help us much, if at all.
That last assertion is not true, in fact. A great deal of exploratory life science research is first accomplished in species like fruit flies, nematode worms, yeast, zebrafish, and the like. Outside the realm of mammals there exists a small menagerie of species that have proven useful in the laboratory. Yet any of that work to ultimately make it to human clinics will first be repeated or confirmed in mammalian species such as mice, dogs, and primates - which might raise the question as to why researchers bother to work with flies, worms, yeast, and fish in the first place.
The answer to that question relates to the bottom line: money, time, resources. Research is by its very nature an exploratory and uncertain business, full of dead ends and unexpected pitfalls. A researcher wants to cover as much ground as he or she can for a given amount of time and money: the more that is explored, the greater the chance of finding a significant path forward. On the one hand, work with mammals will generally produce more useful information, but on the other hand working with mammals, even mice, is very expensive and time-consuming in comparison to working with flies and worms, which in turn is expensive and time-consuming in comparison to working with yeast. If infinite money and time were at hand, all research work would involve mammals, but resources are not infinite and the results of any given study are extremely uncertain.
Bear in mind that evolution produced flies, yeast, fish, and mammals from the same deep roots - and as it turned out, a lot of the mechanisms that link the operation of metabolism with variance of longevity within a species were (a) established very early on, and (b) then didn't change a great deal. It is counter-intuitive to think that researchers can learn useful things about the operation of human biology from yeast (or worms, or flies, or fish), but for some mechanisms and systems they can do just that. The further away from human biology that your model is, the more inference there must be, and the greater the risk that there is in fact some important difference between species that renders your work useless or less valuable - but that doesn't prevent work in lower species from being cost-effective.
So the story is that there is a trade-off in the life sciences between the usefulness of data and the cost of obtaining that data. When you are uncertain of the ultimate value of the work presently being undertaken - i.e. if it seems to have a high risk of failure, or a successful outcome is probably not that valuable in any case - then you won't want to spent much time and money on it until such time as it shapes up. If all indications show a good chance of success and a valuable result, then working with mammals starts to look like a better prospect, however. So we might say that work in flies, worms, yeast, and fish is undertaken in order to justify the cost of exploring the same biology in mammals.