CMScaller: CMS classification

Description Usage Arguments Details Value Note References See Also Examples

View source: R/CMScaller.R

Description

Cancer-cell intrinsic CMS classification based on pre-defined subtype templates.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
CMScaller(
  emat,
  templates = CMScaller::templates.CMS,
  rowNames = "entrez",
  RNAseq = FALSE,
  nPerm = 1000,
  seed = NULL,
  FDR = 0.05,
  doPlot = TRUE,
  verbose = TRUE
)

Arguments

emat

a numeric expression matrix with sample columns, gene rows and Entrez rownames. Microarray data should be normalized. For RNA-seq data, counts or RSEM values could be used directly by setting RNAseq=TRUE.

templates

a data frame with two columns; class (coerced to factor) and probe (coerced to character).

rowNames

a character, either "entrez" (NCBI Entrez), "symbol" (HGNC symbol) or "ensg" (Ensembl). If set to other than "ensg", replaceGeneId is used to translate rownames(emat).

RNAseq

a logical, set to TRUE if emat is untransformed, non-normalized sequencing counts or RSEM values.

nPerm

an integer, number of permutations for p-value estimation.

seed

an integer, for p-value reproducibility.

FDR

a false discovery rate, sets prediction confidence threshold.

doPlot

a logical, whether to produce prediction subHeatmap.

verbose

a logical, whether console messages are to be displayed.

Details

CMScaller provides classification based on pre-defined cancer-cell intrinsic CMS templates. If RNA-seq=TRUE, a pseudocount of 0.25 is added, matrix log2-transformed and quantile normalized (normalizeQuantiles) prior to scaling/centering and prediction. The core algorithm is the Nearest Template Prediction (NTP) algorithm as proposed by Yujin Hoshida (2010). See ntp for further details.

Value

a data frame with class predictions, template distances, p-values and false discovery rate adjusted p-values (p.adjust). Rownames equal emat colnames.

Note

genes with missing values are discarded.

References

Hoshida, Y. (2010). Nearest Template Prediction: A Single-Sample-Based Flexible Class Prediction with Confidence Assessment. PLoS ONE 5, e15543.

Guinney J, Dienstmann R, Wang X, de Reynies A, Schlicker A, Soneson C, et al. The consensus molecular subtypes of colorectal cancer. Nat Med. 2015;21:1350-6.

See Also

ntp, templates.CMS, corCosine

Examples

1
2
3
res <- CMScaller(crcTCGAsubset, RNAseq=TRUE)
head(res)
hist(res$p.value)

peterawe/CMScaller documentation built on June 13, 2020, 4:49 a.m.