BAGS2Clinic: Classification of cell of origin with Nanostring data

In HaemAalborg/hemaClass: One-by-one normalization and classification of hematological malignacies

Description

Classification using the B-cell associated gene signatures (BAGS) classifier of Dybkaer et al. [1] adjusted for nanostring data by Michaelsen et al. [2]. The BAGS classifier attemps to assign the given supplied Diffuse Large B-Cell Lymphomas (DLBCL) samples to one of four classes according to their resembalace to normal Centrocytes (CC), Centroblasts (CB), Memory (M), and Plasmablasts (PB) cells.

Usage

BAGS2Clinic(new.data, cut.spec = NULL, percent.classified = 85)

Arguments

new.data	An expression matrix containing the normalized data, with samples in columns and genes in rows. Rownames should be genenames in hgnc nomenclature.
cut.spec	A numeric value used to threshold the probabilities and determine the class.
percent.classified	A numeric value indicating the percentage of patients that should be classified. An alternative to cut.spec.

Details

The function BAGS2Clinic classifies DLBCL patients according to the cell of origin for the tumor [1,2].

Value

A list of probabilities regarding each patients association with each class, the determined class, and the used cut-off thresholds.

Author(s)

Thomas Yssing Michaelsen <tym (at) bio.aau.dk>
Rasmus Broendum <rfb (at) rn.dk>

References

[1] Dybkaer, Karen, Martin Boegsted, Steffen Falgreen, Julie S. Boedker, Malene K. Kjeldsen, Alexander Schmitz, Anders E. Bilgrau et al. "Diffuse large B-cell lymphoma classification system that associates normal B-cell subset phenotypes with prognosis." Journal of Clinical Oncology 33, no. 12 (2015): 1379-1388.

[2] Michaelsen, T. Y. et al. "A B-cell-associated gene signature classification of diffuse large B-cell lymphoma by NanoString technology." Blood advances, 2(13) (2018): 1542-1546.

Examples

# Nanostring files may be read and normalized using the NanoStringNorm package
library(NanoStringNorm)
data.nano.raw <- read.markup.RCC("path to files")

data.nano.norm <- NanoStringNorm(data.nano.raw, 
                                 Background = "mean",
                                 CodeCount = "sum",
                                 return.matrix.of.endogenous.probes = FALSE)

# Data should be log2 transformed and scaled
data.nano.sc <- microarrayScale(log2(as.matrix(data.nano.norm$normalized.data[,-c(1,2,3)]) + 1))
 
# We may now use the BAGS classifier
BAGS2Clinic(data.nano.sc)
                                                                                                                  

HaemAalborg/hemaClass documentation built on Oct. 22, 2019, 7:01 p.m.