SVA confounding analysis framework

Description Usage Arguments Details Value

Perform SVD confounder analysis and association tests between the Principal Components of the given input values and the phenotype data.

1 2	svd_analysis(values, pdata, center = TRUE, scale = FALSE, rgset = NULL, method = c("lm", "kruskal"))

`values`	A matrix of numerical values. Rows represent variables and columns samples.
`pdata`	A data.frame containing the phenotype data for the samples.
`center`	A logical value indicating whether the variables should be shifted to be zero centered. Alternately, a vector of length equal the number of columns of x can be supplied. The value is passed to scale.
`scale`	A logical value indicating whether the variables should be scaled to have unit variance before the analysis takes place. The default is FALSE for consistency with S, but in general scaling is advisable. Alternatively, a vector of length equal the number of columns of x can be supplied. The value is passed to scale. Use when the variables are in arbitrary units of measurement.
`rgset`	If not NULL, compute association with control probes data contained in the provided RGChannelSet.
`method`	Method used for computing p-values.

This function performs a series of association tests between the Principal Components of the input values and the provided phenotype data. Previously, it centers and scales the data, something that might be useful when the variables are measured in different units.

If a RGChannelSet is given, it can also compute the association with the control probes. This implementation can use two different methods for testing (Linear models and Kruskal-Wallis).

The number of meaningful components is computed using the function EstDimRMT from the package isva.

A list containing the proportion of variance explained by the principal components (variance_explained) and a data.frame representing the results from the association analysis (significance).

Keyeoh/svconfound documentation built on May 15, 2019, 1:25 p.m.