xMLglmnet: Function to integrate predictor matrix in a supervised manner...

In Pi: Leveraging Genetic Evidence to Prioritise Drug Targets at the Gene and Pathway Level

Description

xMLglmnet is supposed to integrate predictor matrix in a supervised manner via machine learning algorithm glmnet. It requires three inputs: 1) Gold Standard Positive (GSP) targets; 2) Gold Standard Negative (GSN) targets; 3) a predictor matrix containing genes in rows and predictors in columns, with their predictive scores inside it. It returns an object of class 'pTarget'.

Usage

xMLglmnet(
df_predictor,
GSP,
GSN,
family = c("binomial", "gaussian"),
type.measure = c("auc", "mse"),
nfold = 3,
alphas = seq(0, 1, 0.1),
standardize = TRUE,
lower.limits = -Inf,
verbose = TRUE,
RData.location = "http://galahad.well.ox.ac.uk/bigdata",
guid = NULL,
...
)

Arguments

df_predictor	a data frame containing genes (in rows) and predictors (in columns), with their predictive scores inside it. This data frame must has gene symbols as row names
GSP	a vector containing Gold Standard Positive (GSP)
GSN	a vector containing Gold Standard Negative (GSN)
family	response family type. It can be one of "binomial" for two-class logistic model or "gaussian" for gaussian model
type.measure	loss to use for cross-validation. It can be one of "auc" for two-class logistic model, "mse" for the deviation from the fitted mean to the response using gaussian model
nfold	an integer specifying the number of folds for cross validataion
alphas	a vector specifying a range of alphas. Alpha is an elasticnet mixing parameter, with 0<=alpha<=1. By default, seq(0,1,by-0.1)
standardize	logical specifying whether to standardise the predictor. If yes (by default), the predictor standardised prior to fitting the model. The coefficients are always returned on the original scale
lower.limits	vector of lower limits for each coefficient (by default, '-Inf'; all should be non-positive). A single value provided will apply to every coefficient
verbose	logical to indicate whether the messages will be displayed in the screen. By default, it sets to TRUE for display
RData.location	the characters to tell the location of built-in RData files. See xRDataLoader for details
guid	a valid (5-character) Global Unique IDentifier for an OSF project. See xRDataLoader for details
...	additional parameters. Please refer to 'glmnet::cv.glmnet' for the complete list.

Value

an object of class "pTarget", a list with following components:

• model: an object of class "cv.glmnet" as a best model

• priority: a data frame of nGene X 5 containing gene priority information, where nGene is the number of genes in the input data frame, and the 5 columns are "GS" (either 'GSP', or 'GSN', or 'NEW'), "name" (gene names), "rank" (ranks of the priority scores), "priority" (priority score; rescaled into the 5-star ratings), and "description" (gene description)

• predictor: a data frame, which is the same as the input data frame but inserting an additional column 'GS' in the first column

• cvm2alpha: a data frame of nAlpha X 2 containing mean cross-validated error, where nAlpha is the number of alpha and the two columns are "min" (lambda.min) and "1se" (lambda.1se)

• nonzero2alpha: a data frame of nAlpha X 2 containing the number of non-zero coefficients, where nAlpha is the number of alpha and the two columns are "min" (lambda.min) and "1se" (lambda.1se)

• importance: a data frame of nPredictor X 1 containing the predictor importance/coefficient info

• performance: a data frame of 1+nPredictor X 2 containing the supervised/predictor performance info predictor importance info, where nPredictor is the number of predictors, two columns are "ROC" (AUC values) and "Fmax" (F-max values)

• gp: a ggplot object for the ROC curve

• call: the call that produced this result

Note

none

See Also

xPredictROCR, xPredictCompare, xSymbol2GeneID

Examples

RData.location <- "http://galahad.well.ox.ac.uk/bigdata"
## Not run: 
pTarget <- xMLglmnet(df_prediction, GSP, GSN)

## End(Not run)

Pi documentation built on Nov. 29, 2021, 3 p.m.