Home

Awesome

Build Status

DLBCLdata

Automated and reproducible download and preprocessing of DLBCL data

The DLBCLdata package for R automates the download and preprocessing of large-scale Gene Expression Profile (GEP) studies of Diffuse Large B-Cell Lymphoma (DLBCL) from the NCBI (National Center for Biotechnical Information) GEO (Gene Expression Omnibus) website. It provides R users with reproducible and easy access to GEP data on GEO as an alternative to the otherwise cumbersome manual downloading and preprocessing. The package handles the RMA preprocessing of the studies of DLBCL using the manufacturer's or custom Brainarray [2] chip definition files (CDF) including the installation of these CDFs.

The package is (hopefully) written with enough generality to allow expansion to other DLBCL and non-DLBCL datasets.

Installation

To install the latest version of DLBCLdata directly from the master branch at GitHub, run

install.packages("devtools")  # If devtools is not installed
devtools::install_github("AEBilgrau/DLBCLdata")

Note, that this version is in development and, as such, it may be unstable. For previous versions of DLBCLdata, visit the old releases at GitHub.

Usage

The package should work with any NCBI GEO repository containing gene expression data. However, the package is tailored specifically to some DLBCL GEO accession numbers. To get an overview of the directly "supported" GEO numbers, see

library("DLBCL_overview"")
View(DLBCL_overview)

To download and process a specific GEO number, GSE56315 [1] say, simply run

res_gse56315 <- downloadAndProcessGEO("GSE56315")

Alternatively, a non-standard CDF file can be specified:

res_gse56315 <- 
  downloadAndProcessGEO("GSE56315", cdf = "brainarray", target = "ensg")

The former downloads the .CEL files and RMA preprocesses the data present in GSE56315 [1] using the standard Affymetrix CDF files. The latter downloads and preprocess directly to Ensembl gene identifiers (ENSG) using RMA normalization and custom Brainarray CDFs [2].

To download and preprocesses all datasets featured in DLBCLdata (shown with DLBCL_overview) using, say, brainarray to Entrez gene identifiers the following line will do so.

dlbcl_data <- downloadAndProcessDLBCL(cdf = "brainarray", target = "entrezg")
str(dlbcl_data, max.level = 2)  # Overview of the object

This function creates the file dlbcl_data.Rds in the working directory which can later be read into R with readRDS.

For more help, see help("DLBCLdata").

References

  1. Dybkaer K, Boegsted M, Falgreen S, Boedker JS 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. (GEO number: GSE56315)

  2. Manhong Dai, Pinglang Wang, Andrew D. Boyd, Georgi Kostov, Brian Athey, Edward G. Jones, William E. Bunney, Richard M. Myer, Terry P. Speed, Huda Akil, Stanley J. Watson and Fan Meng. (2005) "Evolving Gene/Transcript Definitions Significantly Alter the Interpretation of GeneChip Data." Nucleic Acid Research 33 (20), e175 (http://brainarray.mbni.med.umich.edu)

Please also cite DLBCLdata if you use it, see citation("DLBCLdata").