Some excellent answers already to a question I hear frequently. However the question is usually asked to provide the questioner direction on their future career. I used to answer with the 'trend de jour': genomics/transcriptomics/systems biology/synthetic biology/genomics. However in essence the answer should be the same: namely, what do you like doing?
Analysis [any technique esp high throughput]: learn stats, databases and scripting. Also knowing your biology is key and I mean not just the biology of the technique but general biology relating to the disease, trait or structure studied. This is perhaps the biggest field and the most overlooked. Even a basic amount of stats will improve a CV of a candidate looking for work in this area.
Modelling [ systems biology/ population genetics/ ...] Mathematics - inc matlab/octave/R + scripting. Biology stuff as relevant as before
Making stuff - software: [any discipline] learn several languages and understand which libraries will make best use of your dev time. Do not assume that java or perl will be the best language to use for every occasion (mea maxima culpa). What biology to learn will depend on that lab you are in. The lab nearest the cutting edge is best. Howerver the best labs (or company division) may have a project manager to tell you what to do. If you want to be a project manager:- well you should already know the area you are in (I know this is not always the case .. I just feel it should be)
Making stuff - bio-engineering: Call it synthetic biology or what have you. A new field in relation to bioblocks and systems biology, but still in essence the same disciple as the bioreactors/transgenic animals/GM crops of the past 20 years. Creating and or combining bioblocks could be exciting: depending on the project, what luck it has, and its commercial application. Obviously knowing some bench science (bio and/or chem) may help here in addition to at least some basic systems biology depending on your fit to the team/lab that you are in.
Many will see themselves as some of the above. The lack of pigeon holes is a good thing and in general if you are good at several things there is usually room in a group/division to accommodate this (but perhaps more so in academia than industry).