We introduce a model-based hypothesis testing approach for evaluating single-cell RNA-sequencing (scRNA-seq) clusters. This approach is implemented in two ways: (1) a stand-alone clustering pipeline with built-in hypothesis testing to produce clusters corresponding to distinct cell populations and (2) a post-hoc method that can evaluate the statistical significance of any provided set of clusters.
To use the stand-alone clustering pipeline, sc-SHC (single-cell significance of hierarchical clustering), the following command can be used:
source('significance_analysis.R')
clusters <- scSHC(data)
Here, data should be a matrix where the rows are genes and the columns are cells. Optionally, the following parameters can be adjusted:
alpha, which controls the family-wise error rate (default 0.05). If the goal is discovery, consider setting a more lenient alpha, such as 0.25.num_PCs, which controls the number of principal components (default 30).num_features, which controls the number of genes used (default 2500).
To evaluate the significance of any provided set of clusters, the following command can be used:
source('significance_analysis.R')
new_clusters <- testClusters(data,clusters)
Here, data is the same as before, and clusters should be a vector of cluster labels, corresponding to cells in the same order as the columns of the data matrix. The same parameters as above can be adjusted. Additionally, if desired, a given set of genes can be provided through the parameter var.genes rather than allowing our approach to identify informative genes on its own.