21.3 Join tables

We have been working (mostly) with 2 objects so far:

• geneexp that contains gene expression information
• gtf that contains gene annotation information

How can we merge all the information into a single data frame?

tidyr provides an easy way to join 2 data frames, based on one or more columns containing common identifiers, to be able to merge relevant information together.

Relevant functions are the following:

• left_join: keeps all observations from the first table (x)
• right_join: keeps all observations from the second table (y)
• inner_join: keeps the intersection of observations
• outer_join: keeps the union of observations

Let’s try the 4 of them and check how many genes are left in each case (with nrow):

joinL <- left_join(x=gtf, y=geneexp, by=c("gene_symbol" = "GeneSymbol"))
nrow(joinL)

## [1] 420

joinR <- right_join(x=gtf, y=geneexp, by=c("gene_symbol" = "GeneSymbol"))
nrow(joinR)

## [1] 50

joinI <- inner_join(x=gtf, y=geneexp, by=c("gene_symbol" = "GeneSymbol"))
nrow(joinI)

## [1] 49

joinF <- full_join(x=gtf, y=geneexp, by=c("gene_symbol" = "GeneSymbol"))
nrow(joinF)

## [1] 421