jimhester / xenar

Simple R client to UCSC Xena data hubs

title	output	vignette
Simple Client for UCSC Xena Genome Data and Analysis Hub	html_document	% \VignetteIndexEntry{Monday Lab, part B.1: Introduction to Bioconductor} % \VignetteEngine{knitr::rmarkdown}

NOTE: generated from inst/README.Rmd

## Title: Simple Client to UCSC Xena
## Author: Martin Morgan [cre]
## Version: 0.0.1
## 
## -- File: /home/mtmorgan/R/x86_64-unknown-linux-gnu-library/3.2-BiocDevel/XenaR/Meta/package.rds 
## -- Fields read: Title, Author, Version

This package is a simple client to the data management aspects of UCSC Xena. It is very much a work in progress.

Install (one-time; using biocLite() or devtools::install_github()) and then load the library.

source("http://bioconductor.org/biocLite.R")
biocLite("mtmorgan/xenar") 

library(XenaR) 

Discovery

Use XenaHub() to discover available resources, illustrated here exploring available cohorts. It's also possible to explore hosts() and datasets().

xe <- XenaHub()
xe

## class: XenaHub 
## hosts():
##   https://genome-cancer.soe.ucsc.edu/proj/public/xena
## cohorts() (77 total):
##   (unassigned)
##   1000_genomes
##   balagurunathan2008_public
##   ...
##   weir2007_public
##   YauClinical_public
## datasets() (1014 total):
##   public/TCGA/TCGA.UCS.sampleMap/RPPA
##   public/TCGA/TCGA.UCS.sampleMap/HiSeqV2_exon
##   public/TCGA/TCGA.UCS.sampleMap/HiSeqV2
##   ...
##   public/other/1000_genomes/BRCA1
##   public/other/1000_genomes/BRCA2

head(cohorts(xe))

## [1] "(unassigned)"              "1000_genomes"             
## [3] "balagurunathan2008_public" "bardeesy2006_public"      
## [5] "basso2005_public"          "ccle"

Refinement

Having identified one (or more) hosts, cohorts, datasets of interest, provide these as arguments to XenaHub(), e.g., to work with the "ccle" corhort,

xe <- XenaHub(cohorts="ccle")
xe

## class: XenaHub 
## hosts():
##   https://genome-cancer.soe.ucsc.edu/proj/public/xena
## cohorts() (1 total):
##   ccle
## datasets() (4 total):
##   public/other/ccle/CCLE_Expression_2012-04-06.matrix
##   public/other/ccle/ccle_clinicalMatrix
##   public/other/ccle/CCLE_Expression_Entrez_2012-04-06.matrix
##   public/other/ccle/CCLE.copynumber.hugomatrix

datasets(xe)

## [1] "public/other/ccle/CCLE_Expression_2012-04-06.matrix"       
## [2] "public/other/ccle/ccle_clinicalMatrix"                     
## [3] "public/other/ccle/CCLE_Expression_Entrez_2012-04-06.matrix"
## [4] "public/other/ccle/CCLE.copynumber.hugomatrix"

Mix cohorts as required, e.g.,

XenaHub(cohorts=c("ccle", "1000_genomes"))

## class: XenaHub 
## hosts():
##   https://genome-cancer.soe.ucsc.edu/proj/public/xena
## cohorts() (2 total):
##   ccle
##   1000_genomes
## datasets() (6 total):
##   public/other/ccle/CCLE_Expression_2012-04-06.matrix
##   public/other/ccle/ccle_clinicalMatrix
##   public/other/ccle/CCLE_Expression_Entrez_2012-04-06.matrix
##   public/other/ccle/CCLE.copynumber.hugomatrix
##   public/other/1000_genomes/BRCA1
##   public/other/1000_genomes/BRCA2

One is often interested in identifying samples or features present in each data set, or shared by all data sets, or present in any of several data sets. Identifying these samples, including samples in arbitrarily chosen data sets, is shown below.

## samples in each dataset, first host
x <- samples(xe, by="datasets", how="each")[[1]]
lengths(x)        # data sets in ccle cohort on first (only) host

##        public/other/ccle/CCLE_Expression_2012-04-06.matrix 
##                                                        967 
##                      public/other/ccle/ccle_clinicalMatrix 
##                                                        972 
## public/other/ccle/CCLE_Expression_Entrez_2012-04-06.matrix 
##                                                        967 
##               public/other/ccle/CCLE.copynumber.hugomatrix 
##                                                        972

sapply(x, head, 3)                 # 3 sample identifiers

##      public/other/ccle/CCLE_Expression_2012-04-06.matrix
## [1,] "NCIN87_STOMACH"                                   
## [2,] "CMK86_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE"         
## [3,] "PCM6_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE"          
##      public/other/ccle/ccle_clinicalMatrix     
## [1,] "NCIN87_STOMACH"                          
## [2,] "CMK86_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE"
## [3,] "PCM6_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE" 
##      public/other/ccle/CCLE_Expression_Entrez_2012-04-06.matrix
## [1,] "NCIN87_STOMACH"                                          
## [2,] "CMK86_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE"                
## [3,] "PCM6_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE"                 
##      public/other/ccle/CCLE.copynumber.hugomatrix
## [1,] "NCIN87_STOMACH"                            
## [2,] "CMK86_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE"  
## [3,] "PCM6_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE"

## samples common to _all_ datasets, 1st host
y <- samples(xe, by="datasets", how="all")[[1]]
length(y)

## [1] 967

Map(function(x) all(y %in% x), x)

## $`public/other/ccle/CCLE_Expression_2012-04-06.matrix`
## [1] TRUE
## 
## $`public/other/ccle/ccle_clinicalMatrix`
## [1] TRUE
## 
## $`public/other/ccle/CCLE_Expression_Entrez_2012-04-06.matrix`
## [1] TRUE
## 
## $`public/other/ccle/CCLE.copynumber.hugomatrix`
## [1] TRUE

## samples in _any_ dataset, 1st host
z <- samples(xe, by="dataset", how="any")[[1]]
length(z)

## [1] 972

Map(function(x) all(x %in% z), x)

## $`public/other/ccle/CCLE_Expression_2012-04-06.matrix`
## [1] TRUE
## 
## $`public/other/ccle/ccle_clinicalMatrix`
## [1] TRUE
## 
## $`public/other/ccle/CCLE_Expression_Entrez_2012-04-06.matrix`
## [1] TRUE
## 
## $`public/other/ccle/CCLE.copynumber.hugomatrix`
## [1] TRUE

## samples in the first and fourth datasets
datasets(xe)[c(1, 4)]

## [1] "public/other/ccle/CCLE_Expression_2012-04-06.matrix"
## [2] "public/other/ccle/CCLE.copynumber.hugomatrix"

w <- samples(xe, datasets(xe)[c(1, 4)], by="datasets", how="all")[[1]]
length(w)

## [1] 967

head(w)

## [1] "NCIN87_STOMACH"                          
## [2] "CMK86_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE"
## [3] "PCM6_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE" 
## [4] "SNU1105_CENTRAL_NERVOUS_SYSTEM"          
## [5] "DM3_PLEURA"                              
## [6] "NOMO1_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE"

TODO

Lots!

• features() to enable selection of features across datasets, cohorts, etc.
• XenaExperiment() to represent a collection of datasets from XenaHub(), subset to contain specific samples and features.
• Basic data retirieval of all or part of the assays present in a XenaExperiment.