mamba: Fit the MAMBA model.
In dan11mcguire/mamba: MAMBA: Meta-analysis Model Based Assessment of Replicability

Fit the MAMBA model.

1
2
3

mamba(betajk, sjk2, parcores = 1, p = 0.003, lambda = 0.96,
  tau2 = 2e-04, alpha = 3, rel.eps = 1e-08, verbose = 1,
  snpids = NA, maxIter = 10^4L)

`betajk`	Mxk matrix of effect size estimates, where row j corresponds to j-th SNP, and column k corresponds to k-th study . Missing values can be represented with NA.
`sjk2`	Mxk matrix of effect size estimate variances, where row j corresponds to SNP j, and column k corresponds to study k. Missing values can be represented with NA.
`p`	initial value for EM algorithm, for the proportion of non-zero SNPs
`lambda`	initial value for EM algorithm, for the proportion of non-replicable SNPs which are well behaved, or non-outliers.
`tau2`	initial value for EM algorithm, for the variance of replicable non-zero effect SNPs.
`alpha`	initial value for EM algorithm, for the variance inflation of outlier summary statistics at non-replicable SNPs.
`rel.eps`	threshold for when to end EM algorithm. rel.eps = (ll[i] - ll[i-1])/ll[i], where ll[i] is the log-likelihood at iteration i.
`snpids`	an optional vector of SNP id names. If not provided, the ID's will be 1:M, corresponding to the order of the matrix betajk.
`maxIter`	maximum # of EM iterations.