Dear Brian

Essentially, before GWAS the candidate gene approach was the most popular methodology to use.  That said, it hasn't lost all favour - it is now used in conjunction with GWAS to provide more complete pictures for the molecular aetiology of complex traits considering a systems biology approach.

Finding an article comparing things before and after GWAS will be difficult but here are good examples discussing the impact of GWAS in research.

http://onlinelibrary.wiley.com/doi/10.1002/dmrr.1221/abstract;jsessionid=ADC2ED156C67372EB60D14DB57E6613D.f01t04?userIsAuthenticated=false&deniedAccessCustomisedMessage=

http://ajp.psychiatryonline.org/doi/abs/10.1176/appi.ajp.2008.08091354

Regards,

Apart from molecular approaches, purely statistical approaches to genetic questions played a more important role than they do nowadays. For example, BLUP (best linear unbiased prediction) is quite useful to determine the inheritance of complex traits and make predictions on what breeding decisions will increase or decrease the probability of desirable outcomes.

Potential applications of BLUP include complex orthopedic and other diseases in dogs (hip dysplasia, elbow dysplasia, osteochondrosis, epilepsy), but also parameters such as milk production in cows, the career amount of money won in race horses, and so on.