Therefore, alternative approaches have been used. PNA markers are first ordered and their spacings (recombination frequencies) determined by analysis of segregation patterns of the markers within a panel of reference families, which represent finite and limited sample sizes.
Once ordering and spacing are determined between different RFLPs, models are fitted by maximizing the likelihood of the parameters. The likelihoods of different models are compared by calculating the likelihood ratios.
When the likelihood ratio is 1000 : 1(or atleast 100 : 1) then the odds of one model are 1000 (or atleast 100) fold greater than the odds of another, and therefore, the first model is accepted. The 10gIO likelihood ratio is called LOD score (e.g. for 1000 : 1 likelihood ratio, 10g10 = 3, LOD score = 3; for 100 : 1 likelihood ratio, LOD score = 2), and is used for linkage analysis.
A LOD score of 1 to 2 is interesting, 2 to 3 is suggestive and> 3 is proof of linkage. It has been shown that no adjacent pair of markers is separated by more than 32% recombination.



