Selection indices for combining marker genetic data and animal model information

• Main Author: Romano, Eduardo
• Other Authors: Dairy Science
• Format: Others
• Language: en
• Published: Virginia Tech 2014
• Subjects: LD5655.V855 1993.R663; Animal genetic engineering; Animal genetics > Mathematical models; Gene mapping > Evaluation
• Online Access: http://hdl.handle.net/10919/44879; http://scholar.lib.vt.edu/theses/available/etd-09192009-040546/

It was suggested that marker and phenotypic information be combined in order to obtain more accurate or earlier genetic evaluations. An improvement in accuracy or time of evaluation due to utilization of marker assisted selection (MAS) increases genetic progress. Fernando and Grossman (1989) suggested including marker information directly into the Animal Model, Best Linear Unbiased Prediction system, but several problems need to be solved before their approach becomes feasible. Other selection indices were suggested but either do not use all the available information or are suitable only for evaluation of the offspring of the sire from which the marker information was established. A selection index combining marker and Animal Model information was developed to allow comparisons involving offspring, grandoffspring and great-grandoffspring of a sire. Marker information was assumed to be a least square estimate of the difference between the average effects of the two quantitative trait loci (QTL) alleles present in a sire (D_p) and the standard error of this estimate (SE(D_p)). Estimates may have been obtained from a daughter or granddaughter design. Comparisons among grandoffspring and great-grandoffspring also require an estimate of the recombination rate (r) between the marker and the QTL. The Animal Model information consists of predicted transmitting ability (PTA) and reliability of PTA. PTA was assumed not to include any marker information. The expected percentage of the gain in accuracy (PGA) due to the inclusion of marker information in the selection indices is affected by the degree of polymorphism at the marker locus. The polymorphism information content (PIC) of a marker locus was computed for the second and third generations and for mates genotyped or not. PGA increased with larger Dos lower SE(D_p), lower r, a smaller number of own and progeny records, and larger PIC. PGA and PIC reduce over generations. Marker information in dairy cattle is likely to be used in generations beyond offspring. Then, only the use of highly polymorphic markers with a large and accurately estimated effect may be economically justified. === Master of Science