Uses the standard deviation of the difference between each bin to estimate tumor and normal cell clusters. Using method gmm will calculate the per cluster median of the sd, and then fit a two-component GMM to determine tumor cells. If method min_sd is specified (or if only two clusters are provided), simply uses the n_normal_clusts with the lowest median sd.
Usage
identify_normal(
sce,
assay_name,
group_by = "clusters",
method = c("gmm", "min_sd"),
n_normal_clusts = NULL,
plot = TRUE,
use_cnv_score = TRUE
)Arguments
- sce
SingleCellExperiment Object
- assay_name
Name of assay from which to calculate metrics from. It is highly recommended that these are segmented and merged data.
- group_by
Name of column containing the grouping information
- method
One of
gmmormin_sd- n_normal_clusts
Number of expected normal clusters (only for method
min_sd)- plot
Plot cluster scores and tumor/normal identifications
- use_cnv_score
Also use CNV score (absolute mean of the assay)