Predict the parameters or values of additional proteins

This function can either predict the abundance matrix for proteins (type = "response") without missing values according to the linear probabilistic dropout model, fitted with proDA(). Or, it can predict the feature parameters for additional proteins given their abundances including missing values after estimating the hyper-parameters on a dataset with the same sample structure (type = "feature_parameters").

# S4 method for proDAFit
predict(object, newdata, newdesign,
  type = c("response", "feature_parameters"), ...)

Arguments

object	an 'proDAFit' object that is produced by proDA().
newdata	a matrix or a SummarizedExperiment which contains the new abundances for which values are predicted.
newdesign	a formula or design matrix that specifies the new structure that will be fitted
type	either "response" or "feature_parameters". Default: "response"
...	additional parameters for the construction of the 'proDAFit' object.

Value

If type = "response" a matrix with the same dimensions as object. Or, if type = "feature_parameters" a 'proDAFit' object with the same hyper-parameters and column data as object, but new fitted rowData().

Details

Note: this method behaves a little different from what one might expect from the classical predict.lm() function, because object is not just a single set of coefficients for one fit, but many fits (ie. one for each protein) with some more hyper-parameters. The classical predict function predicts the response for new samples. This function does not support this, instead it is useful for getting a matrix without missing values for additional proteins.