Chris White talks a bit about this in his book, IIRC. Apparently it's possible even to differentiate between two different strains of s. cerevisiae in many cases.
The only real way to do it though is to figure it out. Take a known pure sample, plate it, and if all the colonies look the same, you can be confident of what you've got.
Take careful notes of the strain, and perhaps even a photograph. After you do this with enough strains, you can refer to your notes to get a good idea, and eventually you'll even probably be able to at least know whether a given yeast colony is an ale, lager, weizen, or brett strain - or something else entirely - just by looking at it.