DESIGNS OF REFERENCE FAMILIES FOR THE CONSTRUCTION OF GENETIC LINKAGE MAPS

Authors: Da, Yang; Vanraden, Paul; LI, NING - CHINA AGRIC UNIV, BEIJING; WU, CHANGXIN - CHINA AGRIC UNIV, BEIJING

Submitted to: Genetics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/17/1998
Publication Date: N/A
Citation: N/A

Interpretive Summary: Genetic linkage maps of genetic markers are important tools for finding genes associated with economically important traits in domestic animals. The construction of genetic linkage maps requires analyzing genetic marker data generated from individuals in a set of families with designed structure and sufficient DNA resources to generate a large number of genotypes for each individual. Such families are referred to as reference families or a reference family panel. The reference family panel is the foundation of a gene mapping program because it affects cost and quality of genetic linkage maps. The design of the reference families should yield linkage maps with reliable linkage estimates and locus ordering at a minimal gene mapping cost. New developments and findings in this study included the map cost function and ordering reliability as design planning utilities, the optimum detection and ordering levels for various designs, the overlapping of optimum detection level with optimum ordering level, and the map cost and sample size requirements for linkage detection and locus ordering based on the optimal detection and ordering levels. Factors affecting gene mapping designs were evaluated for the first time based on the map cost utility. These developments will aid scientists and animal breeders in designing mating systems that include genetic map information.

Technical Abstract: The reference family panel is the foundation of a gene mapping program because it affects cost and quality of genetic linkage maps. The design of the reference families should yield linkage maps with reliable linkage estimates and locus ordering at a minimal gene mapping cost. A map cost function was defined as the number of genotypes required per marker per unit of map coverage and was used to determine optimal designs. For 6 basic designs, optimum detection level of recombination frequency that minimized map cost was found to be around .11 to .15; corresponding minimal map cost ranged from 10 to 20 genotypes per centi-Morgan coverage per marker, assuming 2 alleles with equal frequency. As family size increases, optimum detection level increases and map cost decreases. Substantial reduction in map cost can be achieved by using highly polymorphic markers. For small families, grandparents are important in reduction of map cost but become less important as family size increases or detection level decreases. Optimum level of recombination frequency for locus ordering was found to be around .12, which overlaps optimum detection level. However, locus ordering would require more genotypes than detecting linkage. For full-sib designs, meioses from both genders should be used for locus ordering. For half-sib designs, reserve families may be needed to order loci that did not have sufficient ordering evidence from reference families. To order closely linked loci, much larger sample size would be required. Sample size and map requirements for various designs were derived; estimates for resources required for locus ordering at optimum detection level and at close linkage as required by positional cloning were provided.