Estimate the expected variance of beta. This is approximately expected(1/var(U)).
expected_vbeta(N0, N1, snps, W, gamma.W, freq, GenoProbList = make_GenoProbList(snps = snps, W = W, freq = freq))
| N0 | The number of Y=0 |
|---|---|
| N1 | The number of Y=1 |
| snps | The snps at which we wish to compute the expected Z Score |
| W | The true causal SNPs (these need not be in "snps") |
| gamma.W | The log odds ratios of effect of the true causal SNPs (not including gamma0, the intercept term) |
| freq | Haplotype frequencies as a data.frame, with column Probability indicating relative frequency in controls. |
| GenoProbList | An list of objects giving the probability of seeing each X,W genotype vector. This can be calculated within the function if no value supplied, or you can pass a pre-calculated version |
The expected variance of beta for each SNP X, assuming the causal SNPs are W
Assumes we have a list, GenoProbList, giving the GenoProb values for each X.
freq=fake_freq(nhaps=100,nsnps=5) # fake haplotype frequency data EVB=expected_vbeta(N0=1000,N1=2000,snps=paste0("s",1:5), W="s1",gamma.W=log(1.5),freq=freq) EVB # causal variant is SNP 1, with OR 1.5#> [1] 0.07152293 0.06514892 0.07261753 0.06969743 0.07155681