rnb.step.betadistribution: rnb.step.betadistribution
Description
Computes the distributions of beta values across various sample groups and adds a corresponding section to the
report.
Usage
rnb.step.betadistribution(rnb.set, report, columns = rnb.getOption("exploratory.columns"), points.per.group = rnb.getOption("distribution.subsample"))
Arguments
rnb.set
HumanMethylation450K dataset as an object of type RnBSet
.
report
Report to contain the methylation deviation section. This must be an object of type
Report
.
columns
Optional; predefined column names (in the form of a character
vector) or indices (an
integer
vector) in the sample annotation table. Only these columns are considered for grouping
samples and defining profiles. All other columns in the phenotype table are ignored.
points.per.group
the targeted number of points (T) per group. Set this to a value < 1 to disable subsampling. More
information in the Details section
Value
The modified report.
Details
If subsampling is enabled (i.e. points.per.group
>0),
observations per group are subsampled according to the following procedure:
Given K groups and numbers of observed beta values per group
N_1,...,N_K, and the target number of points per group T: the total number of points N = sum(N_1,...,N_K) is computed
Afterwards the proportions p_k = N_k/N is computed and from each group, S_k = p_k*(K*T) observations
are randomly selected from all observations belonging to group k.