Tests of candidate genes in breed cross populations for QTL mapping in livestock.

In recent years, several F(2) crosses between outbred lines of livestock have been developed to identify quantitative trait loci (QTL). These populations are valuable for further genetic analysis, including positional candidate gene loci (CGL). Analysis of CGL in F(2) populations is, however, hindered by extensive between-breed linkage disequilibrium (LD). The objectives here were to develop and evaluate three tests for CGL in simulated F(2) breed-cross populations. 1) A standard association test, based on the fixed effect of CGL genotype. This test was significant for CGL at considerable distances from the QTL. 2) A marker-assisted association test, based on a test at the CGL of the fixed effect of CGL genotype in a breed-cross QTL interval mapping model. This removed the impact of between-breed LD, but was not powerful in detecting CGL closely linked to the QTL, unless the CGL was the QTL. 3) An F-drop test, comparing F ratios for a QTL at the CGL with and without the CGL included as fixed effect. It had low power to distinguish close from distant CGL. Power to distinguish two CGL within 10 cM from the QTL was limited and little improved by including QTL effects associated with markers to remove between-breed LD, although the power was greater when one of the CGL was the causative mutation. Therefore, while we conclude that candidate gene tests in QTL mapping populations must be interpreted with caution, we now have a clearer picture of the value of candidate gene tests in these populations.