Linkage analysis has been very successful in identifying genes for many Mendelian diseases, but has not enjoyed the same level of success for complex diseases. A major reason is that complex diseases are multifactorial, involving multiple genes and environmental factors. Linkage analysis is powerful for localizing disease genes with moderate marginal effects with most realistic sample sizes. Traditionally it has been used to search for a single disease locus at one time, and most implementations lack the power to detect genes with small marginal effect but moderate to strong interaction effect with other genes. Thus, methods for detecting gene-gene interaction in linkage analysis are needed. A brief background on gene-gene interaction in linkage analysis is given, with a review of two major approaches: two-locus linkage analysis and interaction analysis. Three methods of interaction analysis are also reviewed: conditional linkage analysis, ordered subset analysis, and generalized estimating equations.