FitLmerContrast - for data with pre post treatment and 2 response groups; within each cell type contrast difference in fold change between groups, baseline difference, and fold change across groups of module scores.

FitLmerContrast - for data with pre post treatment and 2 response groups; within each cell type contrast difference in fold change between groups, baseline difference, and fold change across groups of module scores.

Usage

FitLmerContrast(
  module_data_frame,
  celltype_column = "celltype",
  metadata,
  fixed_effects = NULL,
  lmer_formula = NULL,
  plotdatqc = TRUE,
  figpath
)

Arguments

module_data_frame

data for each cell to model -- designed to be scores for modules (columns) returned by scglmmr::WeightedModuleScore

celltype_column

the column in metadata with the cluster / celltype designation for each cell

lme4metadata

metadata for model fit

f1

model furmula

contrast_list

list of linear model contrasts.

plot_savepath

path to save results

Value

Examples

if (FALSE) {
# load data from single cell data object
Seurat = readRDS("my_seurat_object.rds")

# add cellwise module score for each signature
mod_scores = WeightedCellModuleScore(seurat_object = Seurat,
                                     module_list = btm,
                                     threshold = 0.1,
                                     return_weighted = FALSE, cellwise_scaling = FALSE,
                                     Seurat_version = "2")
Seurat = AddMetaData(Seurat,metadata = mod_scores)
module_n = names(sig_test)

# set up module data frame
module_df = Seurat@meta.data %>% select(barcode_check, celltype_joint, module_n)

# format metadata as factors group_id is order leveled for contrast_fit = contrast(emm1, method = list( (c21 - c20) - (c11 - c10) ))
md = Seurat@meta.data %>%
  mutate(group_id = factor(treat_time,  levels = c('pre_low', 'post_low', 'pre_high', 'post_high'))) %>%
  mutate(sampleid = factor(sampleid)) %>%
  select(barcode_check, celltype_joint, sampleid,  age, group_id)

# Fit mixed model
plot_savepath = paste0(my_figure_save_path, "/marginalmeans/"); dir.create(plot_savepath)

# specify any random intercept model e.g.
f1 = 'modulescore ~ age + group_id + (1|sampleid)'

# fit sc mod mixed model on ewighted module scores.
mm_res = FitLmerContrast(module_data_frame = module_df,
                           celltype_column = 'celltype',
                             metadata = md,
                             fixed_effects = NULL,
                             lmer_formula = f1,
                             plotdatqc = TRUE,
                             figpath = 'your/file/path')
}