Sean Day did research on synesthetically colored letters, which he reported here. What he found was that although there were a few letters where there was a larger group favoring a particular color, this was the exception rather than the rule, and most letters had a fairly even distribution of colors.
Day's results are a little hard to get a sense of from his description; here are the results of another, smaller test (25 subjects), with results that parallel Day's, presented graphically:
In this chart, I've arranged the colors by similarity so that it's easier to see the patterns — or, perhaps more accurately, absence of patterns. For just about every letter, nearly the entire spectrum of colors is represented.
There are some exceptions, though. The letters I, O and X (and also the numerals 0 and 1) were categorized by a majority of synesthetes as uncolored (n/c), white, grey, or black — that is, as colorless. The only letter (or numeral) where there's a striking agreement is the letter A, which (as in Day's study), nearly half the subjects associated with red. Red is the first color that infants can perceive and A is often the first letter that children are taught, which suggests that this correspondence is learned rather than innate.