Phylogenetics of rat inbred strains.

Abstract

Laboratory rats (Rattus norvegicus) have played animportant role in biomedical research for over 100years with more than 200 strains bred as physiologi-cal models of human disease (Lindsey 1979; Steenet al. 1999; Kwitek-Black and Jacob 2001). Unlikemouse inbred strains, little is known about rela-tionships among rat strains (Festing 1979). TheAllele Characterization Project (ACP; lgr.mcw-edu/research/lgr_acp.html) was initiated to guide theidentification of individual strains, ensure purity ofstrains used for specific research purposes, and pro-vide a means of mapping genes based on linkage(Jacob et al. 1995; Brown et al. 1998; Steen et al.1999). The ACP genotyped 48 commonly used ratinbred strains by using over 4800 microsatellitemarkers with known genomic locations. The mi-crosatellite markers are spread relatively evenlyacross the rat genome and have been integrated intothe rat radiation hybrid (RH) map (Steen et al. 1999).Using the ACP microsatellite data and a limitednumber of isozymes, Canzian (1997) conducted agenealogic study of inbred rat strains to better un-derstand relationships among strains and to providea basis for the selection of rat strains for biomedicalresearch. Similar research of mouse inbred strainshas been significant for studies of evolution anddisease (Fitch and Atchley 1985; Beck et al. 2000).The Canzian (1997) study used a simple distancemodel (Fitch and Margoliash 1967) with percentagesof genotype differences. The results of this studyhave been cited by other authors to justify the use ofparticular rat strains for the study of a disease (Rapp2000) and to support the phylogenetic hypothesisthat a given strain of interest is ancestral to otherstrains (D’Cruz et al. 2000; Lorincz et al. 2001).Evolutionary models of microsatellite changehave been shown to more accurately reconstructdivergence patterns involving this type of marker,and there is a rich tradition in the evolutionary bi-ology literature of these models. These models havebeen used to measure population subdivision (Slat-kin 1995), investigate human evolution (Goldsteinet al. 1995b; Perez-Lezaun et al. 1997), map disease-related traits (Fishman et al. 2001), and understandmutations (Colson and Goldstein 1999). Models in-corporating specific hypotheses about mutationalmechanisms of microsatellite markers can clearly besuperior to simple percentage difference models thatpose no hypotheses about how markers mutate. Es-timates of mutation rates for microsatellite markersare approximately 10