benchmark {granulator} | R Documentation |
regress
computes regression between estimated cell
type proportions and the measured cell type proportions (ground truth).
benchmark(deconvoluted, ground_truth)
deconvoluted |
Output object of the function |
ground_truth |
A matrix containing measured cell type proportions in percentages. Samples names are inlcuded in rownames. |
Returns a list containing thres elements:
data: a list of data frames with celltype matched estimated and predicted proportions
stats: a list of data frames with regression statistics comprising Pearson Correlation Coefficient ('pcc'), Concordance Correlation Coefficient ('ccc'), Coefficient of Determination ('adj.r2') and Root Mean Square Error ('rmse')
summary: a data frame with summary statistics by cell type
rank: ranking of deconvolution alghoritms by highest all-to-all correlation of coefficients
summay: summary statistics of regression coefficients by method, signature and cell type
rank: ranking of methods and signatures by highest average regression coefficient
combinations: combination of methods and signatures tested
Vincent Kuettel, Sabina Pfister
# load demo PBMCS data load_ABIS() # deconvolute decon <- deconvolute(m = bulkRNAseq_ABIS, sigMatrix = sigMatrix_ABIS_S0) # bechmark bench <- benchmark(deconvoluted = decon, ground_truth = groundTruth_ABIS)