ON THE USE OF MARKER INFORMATION FROM GRANDDAUGHTER DESIGNS

Selection indices were developed for combining marker information and animal model PTA for offspring, grandoffspring, and great-grandoffspri ng of sires. The sires had marker-QTL linkages identified in granddaug hter designs (GDD). Effects of size and s. e. of marker effects and of weight given to own and progeny records in PTA on gain in accuracy of selection due to a marker were quantified. ie for high weight or larg e s. e., even if the marker effect was substantial. Gains among offspr ing (O), grandoffspring (GO), and great-grandoffspring (GGO), when the marker effect in the sire was the only marker information. Gains decr eased strongly for GO and were smallest for GGO. Assuming that progeny test (PT) daughters of some sons of the sire were also marker genotyp ed, the selection index was modified to incorporate this marker inform ation on sons in addition to that on the sire from the GDD for selecti on of GO. For the current size of the PT, gain from including marker d ata on PT daughters was negligible, requiring 200 to 1000 PT daughters . Polymorphism Information Content (PIC) was computed as the fraction of O, GO, and GGO with marker allelic inheritance traceable to the sir e. PIC in GO exceeded .8 and approached PIC in O only if a marker had at least 10 alleles at equal frequencies.