cruv4_multicov: RUV4's second step.
In dcgerard/vicar: Various Ideas for Confounder Adjustment in Regression

RUV4's second step.

cruv4_multicov(
  Y2,
  alpha,
  sig_diag,
  ctl,
  R22,
  degrees_freedom,
  gls = TRUE,
  likelihood = "t"
)

`Y2`	A matrix of numerics with p rows and q columns, where q is the number of covariates of interest and p is the number of genes.
`alpha`	A matrix of numerics of dimension p by k, where k is the number of hidden confounders. The estimated coefficients of the unobserved confounders.
`sig_diag`	A vector of positive numerics. The estimates of the variances.
`ctl`	A vector of logicals of length `ncol(Y)`. If position i is `TRUE` then position i is considered a negative control.
`R22`	A matrix of numerics numeric. This is the submatrix of the R in the QR decomposition of X that corresponds to the covariates of interest. May be removed in the future.
`degrees_freedom`	A positive numeric. The degrees of freedom of the t-likelihood if using it.
`gls`	A logical. Should we use generalized least squares (`TRUE`) or ordinary least squares (`FALSE`) for estimating the confounders? The OLS version is equivalent to using RUV4 to estimate the confounders.
`likelihood`	Either `"normal"` or `"t"`. If `likelihood = "t"`, then the user may provide the degrees of freedom via `degrees_freedom`.