Affected-Pedigree-Memeber (APM) method P-value?

[D.Nyholt at gu.edu.au]

Hi out there,

Could someone please shed some light on what is the generally accepted 
significant P-value level when using the APM method. I've heard (and 
read) values ranging from 0.00001 (for very complex diseases) to 0.05. 
These values stem from the way in which people treat the APM T-statistic, 
that is, is it a test for linkage or "association". This may seem a 
stupid question however I've seen published signifcance using <0.001 and 
<0.05. Help!

ps. the same question may also be applied to (affected) sib-pair 
analysis.

Any help would be greatly appreciated.

Dale.