PlotModuleTraitCorrelation for one trait

[shiave]

Hi, I'm trying to check one trait (treatment vs control) against the MEs to figure out which modules show strong correlation with the trait. When I try this code:

seurat_obj$group1 <- factor(seurat_obj$group1, levels = c("AV", "MV"))

cur_traits <- c('group1')

seurat_obj <- ModuleTraitCorrelation(
seurat_obj,
traits = cur_traits
)

PlotModuleTraitCorrelation(
seurat_obj,
label = 'fdr',
label_symbol = 'stars',
text_size = 2,
text_digits = 2,
text_color = 'white',
high_color = 'yellow',
mid_color = 'black',
low_color = 'purple',
plot_max = 0.2,
combine=TRUE
)

I get this error:
Error in PlotModuleTraitCorrelation(seurat_obj, label = "fdr", label_symbol = "stars", :
ModuleTraitCorrelation was run only for one trait. Heatmaps are not suggested for visualizing only one variable!

So what is another method I could use to check the correlation?

Some background: We have a treatment group and control group and the seurat object has many brain cell clusters (group column). We only want to look into the microglia group so I ran the following functions on just microglia earlier in my code:

seurat_obj <- SetDatExpr(
seurat_obj,
group_name = "Microglia", # the name of the group of interest in the group.by column
group.by='cluster_id', # the metadata column containing the cell type info.
assay = 'RNA', # using RNA assay
slot = 'counts' # using normalized data
)

compute eigengene-based connectivity (kME):

seurat_obj <- ModuleConnectivity(
seurat_obj,
group.by = 'cluster_id', group_name = 'Microglia'
)

I'm confused about whether this impacts my final correlations when I look at traits versus module eigengene (I don't think it does but I want to make sure).

[smorabit]

Hi,

Based on what you are saying:

Some background: We have a treatment group and control group and the seurat object has many brain cell clusters (group column). We only want to look into the microglia group so I ran the following functions on just microglia earlier in my code:

Instead of module trait correlation I suggest that you run differential module eigengene analysis to compare treatment vs control.