Selection of probes within biofeatures

This step is to select HM450K/EPIC probes, which locate far from TSS (at least 2Kb away) These probes are called distal probes.

Be default, this comprehensive list of TSS annotated by ENSEMBL database, which is programatically accessed using biomaRt to get its last version, will be used to select distal probes. But user can use their own TSS annotation or add features such as H3K27ac ChIP-seq in a certain cell line, to select probes overlapping thoses features regions.

# get distal probes that are 2kb away from TSS on chromosome 1
distal.probes <- get.feature.probe(genome = "hg19", 
                                   met.platform = "450K", 
                                   rm.chr = paste0("chr",c(2:22,"X","Y")))

Creation of a MAE object

library(MultiAssayExperiment)
library(ELMER.data)
data(LUSC_RNA_refined,package = "ELMER.data")
data(LUSC_meth_refined,package = "ELMER.data")
GeneExp[1:5,1:5]

##                 TCGA-22-5472-01A-01R-1635-07 TCGA-22-5489-01A-01R-1635-07
## ENSG00000188984                    0.0000000                     0.000000
## ENSG00000204518                    0.4303923                     0.000000
## ENSG00000108270                   10.0817831                    10.717673
## ENSG00000198691                    6.4462711                     6.386644
## ENSG00000135776                    8.5929182                     9.333097
##                 TCGA-22-5491-11A-01R-1858-07 TCGA-56-5898-01A-11R-1635-07
## ENSG00000188984                     0.000000                    0.5233612
## ENSG00000204518                     0.000000                    0.0000000
## ENSG00000108270                    10.180863                   10.1595162
## ENSG00000198691                     5.755627                    4.8354795
## ENSG00000135776                     8.558358                    8.7772810
##                 TCGA-90-6837-01A-11R-1949-07
## ENSG00000188984                     0.000000
## ENSG00000204518                     1.167294
## ENSG00000108270                     9.975092
## ENSG00000198691                     3.152329
## ENSG00000135776                     8.604804

Meth[1:5,1:5]

##            TCGA-43-3394-11A-01D-1551-05 TCGA-43-3920-11B-01D-1551-05
## cg00045114                    0.8190894                    0.8073763
## cg00050294                    0.8423084                    0.8241138
## cg00066722                    0.9101127                    0.9162212
## cg00093522                    0.8751903                    0.8864599
## cg00107046                    0.3326016                    0.3445508
##            TCGA-56-8305-01A-11D-2294-05 TCGA-56-8307-01A-11D-2294-05
## cg00045114                    0.8907009                    0.8483227
## cg00050294                    0.5597787                    0.3488952
## cg00066722                    0.7228874                    0.6238963
## cg00093522                    0.8050060                    0.8194921
## cg00107046                    0.4312738                    0.4328108
##            TCGA-56-8308-01A-11D-2294-05
## cg00045114                    0.7612094
## cg00050294                    0.3908054
## cg00066722                    0.7727631
## cg00093522                    0.7507631
## cg00107046                    0.4260053

mae <- createMAE(exp = GeneExp, 
                  met = Meth,
                  save = TRUE,
                  linearize.exp = TRUE,
                  save.filename = "mae.rda",
                  filter.probes = distal.probes,
                  met.platform = "450K",
                  genome = "hg19",
                  TCGA = TRUE)
as.data.frame(colData(mae)[1:5,])  %>% datatable(options = list(scrollX = TRUE))

as.data.frame(sampleMap(mae)[1:5,])  %>% datatable(options = list(scrollX = TRUE))

as.data.frame(assay(getMet(mae)[1:5,1:5]))  %>% datatable(options = list(scrollX = TRUE))

as.data.frame(assay(getMet(mae)[1:5,1:5])) %>% datatable(options = list(scrollX = TRUE))

Using non-TCGA data

In case you are using non-TCGA data there are two matrices to be inputed, colData with the samples metadata and sampleMap, mapping for each column of the gene expression and DNA methylation matrices to samples. An simple example is below if the columns of the matrices have the same name.

library(ELMER)
# example input
met <- matrix(rep(0,15),ncol = 5)
colnames(met) <- c("Sample1",
                   "Sample2", 
                   "Sample3",
                   "Sample4",
                   "Sample5")
rownames(met) <- c("cg26928153","cg16269199","cg13869341")

exp <- matrix(rep(0,15),ncol = 5)
colnames(exp) <- c("Sample1",
                   "Sample2", 
                   "Sample3",
                   "Sample4",
                   "Sample5")
rownames(exp) <- c("ENSG00000073282","ENSG00000078900","ENSG00000141510")


assay <- c(rep("DNA methylation", ncol(met)),
           rep("Gene expression", ncol(exp)))
primary <- c(colnames(met),colnames(exp))
colname <- c(colnames(met),colnames(exp))
sampleMap <- data.frame(assay,primary,colname)

distal.probes <- get.feature.probe(genome = "hg19", 
                                   met.platform = "EPIC")

colData <- data.frame(sample = colnames(met))
rownames(colData) <- colnames(met)

mae <- createMAE(exp = exp, 
                 met = met,
                 save = TRUE,
                 filter.probes = distal.probes,
                 colData = colData,
                 sampleMap = sampleMap,
                 linearize.exp = TRUE,
                 save.filename = "mae.rda",
                 met.platform = "EPIC",
                 genome = "hg19",
                 TCGA = FALSE)