CERVUS3.0说明书(20)

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。

The proportion of candidate parents sampled sets an overall ceiling on the success rate of parentage analysis. In addition, critical LOD or Delta scores increase as the

proportion of candidate parents sampled declines, further diminishing the success rate of parentage analysis. If the proportion of candidate parents sampled is less than 0.5, the success rate tends to be low unless loci are very highly polymorphic and/or the total number of candidate parents is small.

If you don't have samples from any candidate parents Cervus is not the appropriate software. Instead you should do a sibship analysis.

Analysing parentage of sib groups

In many animals and plants, offspring can be grouped into sibships. For seeds in a pod the question of interest may be: "Were all the seeds pollinated by the same plant?" The philosophy of Cervus is to make as few assumptions as possible when evaluating parentage. In the context of sibs, that means evaluating parentage of each sib

independently. In other words, list sibs in the offspring file in exactly the same way as you would list unrelated offspring and a separate parentage test will be carried out for each.

To answer questions about multiple paternity, for example, you need a decision rule. The null hypothesis for each sibship is that paternity is attributable to a single male, and this can be rejected in favour of multiple paternity if, for example, at least two offspring within a sibship have paternity assigned with 95% confidence to different males.

Cervus may be used to eliminate all sampled candidate parents with 95% confidence, so that for example extra-pair paternity can be inferred even when the true father is not sampled.