CERVUS3.0说明书(19)

parentage analysis program整理自软件help文件,下载过的朋友可以直接在help里查看。

This set of twelve loci is sufficient to assign parentage with 80% confidence to the

offspring of sampled mothers but is insufficient to assign parentage with 95% confidence to the offspring of sampled mothers or to assign parentage even at 80% confidence to the offspring of unsampled mothers.

The simulation accurately predicts the success of paternity analysis with the real data. The corrected likelihood equations of Kalinowski et al. (2007) significantly increase the

number of paternities assigned at both confidence levels when mothers are sampled and when mothers are unsampled.

To further improve the success of paternity assignment at strict confidence or for

offspring with unsampled mothers it is necessary to type additional loci.

The ongoing study in Rum red deer now uses a different panel of markers for

paternity analysis. The three allozyme loci (which provided little statistical power for the effort involved in typing them) have been discarded and six additional

microsatellite loci have been added, giving a total of fifteen microsatellite loci. Both for offspring with sampled mothers and offspring with unsampled mothers, this set of markers has sufficient power to either assign paternity with 95% confidence or to make no assignment. Because the power of the analysis is high it is reasonable to assume that offspring not assigned paternity were sired by unsampled candidate fathers.

Unsampled candidate parents

Cervus can only assign parentage to candidate parents who have been sampled and typed. However, you can take account of the presence of unsampled candidate parents in Cervus simulations by setting the simulation parameter proportion of candidate parents sampled to a value less than 1. Then during parentage analysis, Cervus can estimate confidence of each parentage assignment taking account of the possibility that an unsampled candidate parent might be the true parent.