Calculate VISION pathway scores with auto-correlation (for meta cells) — rs_mc_vision_with

The function will take in a list of gene sets that contains lists of "pos" and "neg" gene indices (0-indexed). You don't have to provide the "neg", but it can be useful to classify the delta of two stats (EMT, Th1; Th2) etc. Additionally, it will take a random gene list and calculate an auto-correlation score based on Geary's C to identify pathways that show significant patterns on the kNN graph generated on the provided embedding. This version works on MetaCell counts which are stored in memory directly.

Usage

rs_mc_vision_with_autocorrelation(
  sparse_data,
  embd,
  knn_data,
  gs_list,
  random_gs_list,
  vision_params,
  cluster_membership,
  verbose,
  seed
)

Arguments

sparse_data: A named list that needs to have data, indptr, indices, nrow, ncol and cs_type. Shape is (metacells, genes) and the data need to be the normalised counts.
embd: Numerical matrix. The embedding matrix to use to generate the kNN graph. Needs to be of the same order/length as the meta cells in sparse_data.
knn_data: Optional list. This contains pre-computed kNN data (indices, dist, k) and the dist_metric it was built with. The user has to ensure consistency! If provided, this will be used rather than a graph built from the parameter list.
gs_list: Nested list. Each sublist contains the (0-indexed!) positive and negative gene indices of that specific gene set.
random_gs_list: Double-nested list. The outer list represents the clusters of gene sets and the inner list represents the permutations within that cluster.
vision_params: List. The kNN parameters, only read when no knn_data is provided.
cluster_membership: Integer vector. 1-indexed(!) position in random_gs_list of the permuted cluster each gene set belongs to.
verbose: Integer. 0L - quiet; 1L - normal verbosity; 2L - detailed verbosity.
seed: Integer. Random seed for reproducibility.

Value

A list with the following items:

autocor_res - List with auto_cor (1 - Geary's C), p_val and fdr, one entry per gene set.
vision_mat - A matrix of meta cells x vision scores per gene set.

References

DeTomaso, et al., Nat. Commun., 2019