As discussed in #141 and #142, there's a new RNGseed argument to BiocParallel param objects. knntlOptimisation() has a seed parameter that probably should now be somehow passed to the parallelisation parameter passed as BPPARAM (if not already set). The exact way to tackle this should probably be discussed further, but I suspect that the way the random seed is handled is not valid anymore.
As discussed in #141 and #142, there's a new
RNGseedargument to BiocParallel param objects.knntlOptimisation()has aseedparameter that probably should now be somehow passed to the parallelisation parameter passed asBPPARAM(if not already set). The exact way to tackle this should probably be discussed further, but I suspect that the way the random seed is handled is not valid anymore.cc @lmsimp