summarizeReplicates: Summarizes between normalized replicate values given in a...

In cellHTS: Analysis of cell-based screens

Description

Summarizes the normalized (and possibly already scored) replicate values given in a cellHTS object, and calculates a single z-score value for each probe.

Currently this function is implemented only for single-color data.

Usage

summarizeReplicates(x, zscore, summary = "min")

Arguments

x	a cellHTS object that has already been normalized (see details).
zscore	indicates if the replicate values should be centered and scaled. If missing (default), the data will not be centered and scaled. Otherwise, the value of this argument should be a character string, either "+" or "-", specifying the sign to use for the calculated z-scores (see details).
summary	a character string indicating how to summarize between replicated measurements. One of "min" (default), "mean", "max", "rms", "closestToZero", or "FurthestFromZero" can be used (see details).

Details

Given the normalized values given in the slot xnorm of x, a single z-score is calculated for each probe.

The argument zscore indicates the state of the normalized replicate measurements: if zscore is missing, it is assumed that the replicates have been scored, by calling normalizePlates with the argument zscore equal to "-" or "+"; Otherwise, zscore should be given, so that a robust z-score is calculated for each plate and each well by subtracting the overall median and dividing by the overall mad. The overall median and mad are taken by considering the distribution of intensities (over all plates) in the wells whose content is annotated as sample. The allowed values for zscore ("+" or "-") are used to set the sign of the calculated z-scores. For example, with a zscore="-" a strong decrease in the signal will be represented by a positive z-score, whereas setting zscore="+", such a phenotype will be represented by a negative z-score. This option can be set to calculate the results to the commonly used convention.

Finally, a single z-score per probe is calculated by summarizing between scored replicates. If summary="mean", the average of replicate values is considered; if summary="max", then the maximum of replicate intensities is taken; if summary="min", the minimum is considered, instead (conservative); if summary="rms", the square root of the mean squared value of the replicates (root mean square) is taken as a summary function; if summary="closestToZero", the value closest to zero is taken as a summary (ueful when both sides of the distribution of z-score values are of interest); if summary="furthestFromZero", the value furthest from zero is taken as a summary (ueful when both sides of the distribution of z-score values are of interest)

Value

An object of class cellHTS, which is a copy of the argument x plus the slot score, a numeric vector containing the z-factor for each well in every plate. The length of this vector is therefore equal to the product between the plateSize and the number of plates. Moreover, the processing status of the cellHTS object is updated in the slot state to state["scored"]= TRUE.

Author(s)

W. Huber [e-mail], Ligia Braz [e-mail]

See Also

normalizePlates, summarizeChannels

Examples

    data(KcViabSmall)
    x <- normalizePlates(KcViabSmall, normalizationMethod="median")
    x <- summarizeReplicates(x, zscore="-", summary="min")

cellHTS documentation built on Oct. 5, 2016, 4:40 a.m.