NormalizeBeta {MAGeCKFlute} | R Documentation |
Two normalization methods are available. cell_cycle
method normalizes gene beta scores
based on positive control genes in CRISPR screening. loess
method normalizes gene
beta scores using loess.
NormalizeBeta(beta, samples = NULL, method = "cell_cycle", posControl = NULL, minus = 0.2)
beta |
Data frame, in which rows are EntrezID, columns are samples. |
samples |
Character vector, specifying the samples in |
method |
Character, one of 'cell_cycle'(default) and 'loess'. |
posControl |
A file path or a character vector, specifying a list of gene entrezids as positive controls used for cell cycle normalization |
minus |
Numeric, scale for cell cycle normalization. Between 0 and 1. |
In CRISPR screens, cells treated with different conditions (e.g., with or without
drug) may have different proliferation rates. So we defined a list of core essential genes,
which is equally negatively selected between samples with different proliferation rate.
Normalization of gene beta scores is performed using these essential genes. cell_cycle
in MAGeCKFlute normalizes the beta scores of all genes based on the median beta score of essential genes.
After normalization, the beta scores are comparable across samples. loess
is another
optional normalization method, which is used to normalize array data before.
A data frame with same format as input data beta
.
Wubing Zhang
data(MLE_Data) # Read beta score from gene summary table in MAGeCK MLE results dd = ReadBeta(MLE_Data, organism="hsa") tmp = TransGeneID(rownames(dd), "Symbol", "Entrez") dd = dd[!(duplicated(tmp)|is.na(tmp)), ] rownames(dd) = tmp[!(duplicated(tmp)|is.na(tmp))] samples=c("D7_R1", "D7_R2", "PLX7_R1", "PLX7_R2") #Cell Cycle normalization dd_essential = NormalizeBeta(dd, samples=samples, method="cell_cycle") head(dd_essential) #Optional loess normalization dd_loess = NormalizeBeta(dd, samples=samples, method="loess") head(dd_loess)