plotRankedFeature: Plot top outcome-associated features

In BioMM: BioMM: Biological-informed Multi-stage Machine learning framework for phenotype prediction using omics data

Description

Plot top ranked outcome-associated features from stage-2 data. The ranking criteria are based on metrics such as Nagelkerke pseudo R-square.

Usage

plotRankedFeature(
  data,
  posF = TRUE,
  topF = 10,
  blocklist,
  binarize = FALSE,
  rankMetric = c("AUC", "R2", "Zscore", "negPlogit", "negPwilcox", "size"),
  colorMetric = c("AUC", "R2", "Zscore", "negPlogit", "negPwilcox", "size"),
  core = MulticoreParam(),
  pathTitle = "GO pathways",
  fileName = NULL
)

Arguments

data	The input stage-2 data (either data.frame or matrix). Rows are the samples, columns are pathway names, except that the first column is the label (the outcome).
posF	A logical value indicating if only positively outcome-associated features should be used. (Default: TRUE)
topF	The top ranked number of features at stage-2 (topF >= 2). (Default: 10)
blocklist	A list of matrices with block IDs as the associated list member names. The block IDs identical to the stage-2 feature names. For each matrix, rows are the samples and columns are the probe names, except that the first column is named 'label'. See also omics2pathlist.
binarize	A logical value indicating if the individual features under investigation should be binarized. The default is FALSE, which provides the estimated class probabilities for each pathway-level feature. If TRUE, then the binary output is given for each feature.
rankMetric	A string representing the metrics used for ranking. Valid options are c("AUC", "R2", "Zscore", "negPlogit", "negPwilcox"). "negPlogit" denotes the negative log P value from the logistic regression and "negPwilcox" means the negative log P value based on the Wilcoxon test. "size" is the block size.
colorMetric	A string representing the metric used to color the plot. Valid options are c("AUC", "R2", "Zscore", "negPlogit", "negPwilcox"). "negPlogit" denotes the negative log P value from the logistic regression and "negPwilcox" means the negative log P value based on wilcoxon test. "size" is the block size.
core	The number of cores used for computation. (Default: 10)
pathTitle	A string indicating the name of pathway under investigation.
fileName	The plot file name. (Default: 'plottopF.png')

Details

If the argument posF is TRUE, and no positively outcome-associated features are present in stage-2 data , then an error is reported. In addition, if topF is bigger than the number of positively outcome-associated features, an error is returned.

Value

An output image file and the summary statistics of the top pathways.

References

Perlich, C., & Swirszcz, G. (2011). On cross-validation and stacking: Building seemingly predictive models on random data. ACM SIGKDD Explorations Newsletter, 12(2), 11-15.

See Also

omics2pathlist.

BioMM documentation built on Nov. 8, 2020, 11:04 p.m.