I am doing my PhD in algebraic graph theory, for not much more reason than that was what was available. However, I love deep structure and theory in mathematics, and I do not particularly want to be a graph theorist for the rest of my life.

I have heard of mathematicians changing from more theoretical subjects to broad ones like combinatorics, but not the other way round, and am concerned that this is because the time it takes to learn the requisite theory for deeper subjects is hard to come by after grad school.

Has anyone done this, or known of someone who has? Is it at all likely or possible I will be able to get a post-doc position in a subject only marginally connected to my thesis? I have been using small amounts of algebraic number theory, and would ideally like to go on and specialise in something similar.

Any advice much appreciated!

EDIT (June '12) - I came across this old question of mine, and now feel rather silly for having asked it at all. For the record, and for anyone who might be having similar doubts to me: a year or so on I have realised that, as was mentioned below, I should not pigeonhole myself, and that the most interesting problems are often those that lead to other seemingly disparate subfields. Perhaps more to the point, I actually can't now think of a subject I'd rather do than algebraic graph theory! If you find yourself in the position I was in, I advise you to just look for those problems you find most interesting in your "chosen" field - they are bound to lead to other areas. And anyway, surely all that matters is that you are interested and inspired?! In my own research I have used number theory, galois theory, group theory, and even a bit of probability. I might even opine that graph theory is one of the subjects most likely to appear in intra-disciplinary work, which seems to be forming an ever higher proportion of mathematical research.