Skip to contents

Compare clusters using hypergeometric test

Source: R/hc_cluster_compare.R
hc_cluster_compare.Rd

hc_cluster_compare() compares clusters from two AnnDatR objects using a hypergeometric test to assess the significance of overlap between clusters.

Usage

hc_cluster_compare(
  AnnDatR_A,
  AnnDatR_B,
  graph_type = c("bipartite", "optimized")
)

Arguments

AnnDatR_A

First AnnDatR object

AnnDatR_B

Second AnnDatR object

graph_type

Type of graph to return: "bipartite" for simple overlap graph, "optimized" for optimized matching graph (default: "bipartite")

Value

A list containing:

  • matches: A tibble of cluster comparisons with overlap statistics and p-values.

  • heatmap: A ggplot2 heatmap visualizing cluster overlaps.

  • match_score (if graph_type is "optimized"): A numeric score representing the total normalized overlap percentage.

Details

If graph_type is set to "optimized", the function will run a cluster matching algorithm to ensure no many-to-many matches between clusters, while still allowing one-to-many or many-to-one matches. The algorithm will:

  1. Filter significant matches based on adjusted p-value threshold (0.05)

  2. Sort matches by overlap size in descending order

  3. Iteratively add matches to the result set, ensuring no many-to-many matches occur.

  4. Calculate a match score representing the total normalized overlap percentage.

Examples

# Run clustering A
adata_res <- hc_pca(example_adata, components = 40)
adata_res <- hc_distance(adata_res, components = 20)
adata_res <- hc_snn(adata_res, neighbors = 15)
#> Building SNN based on a provided distance matrix
#> Computing SNN
adata_res_a <- hc_cluster_consensus(adata_res, resolution = 9)
#> Iteration: 0 *** value: 951.642
#> Iteration: 1 *** value: 198.957
#> Iteration: 2 *** value: 73.707
#> Iteration: 3 *** value: 64.7548
#> Iteration: 4 *** value: 47.8331
#> Iteration: 5 *** value: 47.6047
#> Iteration: 6 *** value: 46.7243
#> Iteration: 7 *** value: 46.1215
#> Iteration: 8 *** value: 46.1215
#> Minimum: 46.1215
#> Joining with `by = join_by(cons_cluster)`

# Run clustering B
adata_res_b <- hc_cluster_consensus(adata_res, resolution = 7)
#> Iteration: 0 *** value: 948.279
#> Iteration: 1 *** value: 70.5063
#> Iteration: 2 *** value: 22.246
#> Iteration: 3 *** value: 22.2447
#> Iteration: 4 *** value: 22.2447
#> Minimum: 22.2447
#> Joining with `by = join_by(cons_cluster)`

# Compare clusters between the two results
hc_cluster_compare(adata_res_a, adata_res_b)
#> 0.00% of genes do not match between the two datasets.
#> Warning: The `size` argument of `element_line()` is deprecated as of ggplot2 3.4.0.
#> ℹ Please use the `linewidth` argument instead.
#> ℹ The deprecated feature was likely used in the HPAclusteR package.
#>   Please report the issue at
#>   <https://github.com/buenoalvezm/HPAclusteR/issues>.
#> $matches
#> # A tibble: 42 × 8
#>    cluster_A cluster_B n_genes_A n_genes_B n_overlap percentage_overlap    p_val
#>    <chr>     <chr>         <int>     <int>     <int>              <dbl>    <dbl>
#>  1 1         1                36        45        36              100   1.26e-57
#>  2 10        10               19        18        18              100   2.01e-37
#>  3 11        11               28        30        28              100   3.35e-52
#>  4 12        4                21        19        18               94.7 2.67e-34
#>  5 13        13               12         7         7              100   4.66e-15
#>  6 13        3                12        45         4               33.3 1.46e- 3
#>  7 14        16               24        21        21              100   1.92e-40
#>  8 15        18               26        27        26              100   2.50e-50
#>  9 16        19                7        26         4               57.1 1.29e- 5
#> 10 16        27                7        33         2               28.6 2.07e- 2
#> # ℹ 32 more rows
#> # ℹ 1 more variable: adj_p_val <dbl>
#> 
#> $heatmap

#> 
#> $network

#>