R/within_data_correlation.r
within_data_correlation.RdCreate correlation plots of the effect sizes, between random permutations and subsets of a given dataset
within_data_correlation(
data,
output_path = getwd(),
sampleID = "donor_id",
design = "placeholder",
sexID = "sex",
celltypeID = "cell_type",
coeff = "male",
N_randperms = 5,
N_subsetpairs = 5,
y = NULL,
region = "single_region",
control = NULL,
pval_adjust_method = "BH",
rmv_zero_count_genes = TRUE
)the input data (should be an SCE object)
base path in which outputs will be stored
sample ID
the design formula of class type formula. Equation used to fit the model- data for the generalised linear model e.g. expression ~ sex + pmi + disease
sex ID
cell type ID
which coefficient to carry out DE analysis with respect to
number of randomised permutations of the dataset (based on sex) to be correlated
number of pairs of subsets of the dataset to be correlated
the column name in the SCE object for the return variable e.g. "diagnosis" - Case or disease. Default is the last variable in the design formula. y can be discrete (logistic regression) or continuous (linear regression)
the column name in the SCE object for the study region. If there are multiple regions in the study (for example two brain regions). Pseudobulk values can be derived separately. Default is "single_region" which will not split by region.
character specifying which control level for the differential expression analysis e.g. in a case/control/other study use "control" in the y column to compare against. NOTE only need to specify if more than two groups in y, leave as default value for two groups or continuous y. Default is NULL.
the adjustment method for the p-value in the differential expression analysis. Default is benjamini hochberg "BH". See stats::p.adjust for available options
whether genes with no count values in any cell should be removed. Default is TRUE Saves all plots in the appropriate directory