Recombinant-inbred strains: general methodological considerations relevant to the study of complex characters

Abstract

If appropriately determined, recombinant-inbred (RI) strain means provide an excellent method for determining genetic correlations among complex characters. However, little systematic attention has been paid to important environmental influences on strain means such as random effects due to litter membership or systematic maternal influences, which are inevitably confounded with genetic effects. It is suggested that users of RI strains would do well to control for litter effects by sampling appropriately from many litters and assessing the potential role of maternal influences by appropriate fostering procedures. Concern for these and other environmental sources of variation has caused reliability of strain means to emerge as an important issue in studies with RIs which focus on complex characters. Examples of estimating the reliability of RI strain means are provided to draw attention to the value of this kind of information in both gene-mapping studies and genetic correlational analyses. In addition, particularly in the case of behavioral tests which are susceptible to considerable day-to-day variation, repeated testing of the same animals can serve to diminish the influence of extreme deviations which are due to random variations in the manner in which the test is conducted on any given day. The advantages of RIs for gene mapping are well established. However, via the power of genetic correlational analysis, the RI methodology is emerging as a major alternative method, e.g., as distinct from lesion studies, pharmacological interventions, etc., in the bio-behavioral sciences to explore relationships between different domains of inquiry. Via its cumulative and integrative power, it is likely to make a major contribution to investigations of relationships between complex characters at various levels of and this application which should be considered separately from its application to gene mapping.