The often seen in GWAS studies *p*-value is a measure of the probability that a suspect association is purely by chance, and not a real causal link. More specifically, it is the probability of obtaining equal or stronger evidence in favor of the linkage in presence of the null hypothesis (that there is no linkage) than what was actually obtained. This, and other assumptions built into it, make it complex and tricky measure to use in practice.

In contrast, posterior probability of linkage seeks to directly determine the probability that the linkage is real. It's proponents believe it is less ambiguous giving the probability they actually want to know, and it is more readily adaptable to analysis of complex inter-related pedigrees. In particular it's not a reverse of the *p*-value, in other words *p*-value plus PPL do not equal one, nor can one be obtained from another, nor same thresholds applied due to different assumptions and meaning. A value over 0.02 or 2% is considered evidence towards linkage and below it against linkage, although higher thresholds such as 50% or 75% are frequently employed to prioritize further study. A slight variation of the algorithm is PPLD, posterior probability of LD. For PPLD the base probability dividing evidence for and against is 0.004.

PPL and PPLD are chiefly analyzed with statistical genetic software package KELVIN [PMID 22189470].