Is there still a need for candidate gene approaches in the era of genome-wide association studies?

Is there still a need for candidate gene approaches in the era of genome-wide association studies?