Molecular genetics significantly contributed to the current concepts of the etiology of Alzheimer disease (AD). The genetic association between APOE epsilon4 and both sporadic and familial late-onset AD (LOAD) was discovered almost one decade ago. Soon after this breakthrough it became clear that there should exist additional risk alleles of other genes in order to fully explain the proportion of AD attributable to genetic factors. However, up to now none of the numerous studies involving more than 100 candidate genes revealed convincing evidence for any predisposing risk alleles in genes other than APOE. This review briefly discusses possible reasons for this lack of success and proposes criteria for a more efficient selection of positional and functional candidate genes for LOAD.