Genomewide studies of genetic association may prove less useful for individual risk estimation than for discovering and characterizing pathogenetic molecular pathways (Hirschhorn, 2009). In the latter task, they complement the traditional tools of bioinformatics and experimental biology. They contribute the empirical muscle that only the epidemiological observation of the disease process in humans may provide. This chapter illustrates this with the aid of a study of the types of ion channel dysfunction that may cause multiple sclerosis (MS). From a methodological point of view, this chapter complements Chapter 14 in this volume by tackling a form of interaction where the value of a causal factor determines a change of direction in the effect of another factor (Azzalini and Cox, 1984). This form of interaction, known as qualitative, is interpretable in terms of mechanism. We propose a method for testing for qualitative interaction between genetic effects, in the context of the analysis of family structured genetic data. We illustrate the method with the aid of the above mentioned study.