Presto for objects leveraging HDF5 storage

[wish1832]

Hello,

I have a very large single cell dataset that must be saved using HDF5 storage due to high RAM usage. To do this, I'm using the HDF5Array package with a SingleCellExperiment object, which saves expression matrices as DelayedArray objects. When I attempt to run presto on an object with DelayedArray matrices, I receive the following error:

 wilcoxauc(
 +   sce,
 +   assay = "logcounts",
 +   group_by = "clusters"
 + )

Error in UseMethod("rank_matrix") :
no applicable method for 'rank_matrix' applied to an object of class "c('DelayedMatrix', 'DelayedArray', 'DelayedUnaryIsoOp', 'DelayedUnaryOp', 'DelayedOp', 'Array', 'RectangularData')"

Is there a plan to extend presto to work with DelayedArray matrices? Seurat has also recently implemented HDF5 storage with the BPCells package, so it would also be useful to have presto be able to work with Seurat/BPCells as single cell datasets continue to scale upward in size.