Single-Nucleotide Polymorphisms in case-control association studies: candidate genes or whole genome