I am trying to find the likelihood ratio for a 2 x3 table (actually SNP data). I do not want the log likelihood. Is this the correct R code for the following table? Row totals are fixed at 100 each (case control situation). aa Aa AA Total 10 20 70 100 20 40 50 100 30 60 120 So the row totals are fixed. Is this the correct code3? Likelihood.ratio dmulitnom(100, c(0.1, 0.2, 0.7))*dmultinom(100, c(0.2, 0.4, 0.5)) divided by dmulitnom(100, c(0.33, 0.33, 0.34))*dmultinom(100, c(0.33, 0.33, 0.34)) What about if the null came from the HWE proportions how will this work for the case control SNP situation? -- Thanks, Jim. [[alternative HTML version deleted]]