Hi Community,

I've been working through the OSCA workflow for single cell Lun 416B cell line (http://bioconductor.org/books/3.14/OSCA.workflows/lun-416b-cell-line-smart-seq2.html).

I'm trying to understand the logic of what is happening in the argument, "col", in the plot below as I haven't come across parentheses followed by brackets used in this way and couldn't insert an operator in between that returned the same result.

The individual components make sense; c("black", "red") is specifying two colours. grepl() is a search through "phenotype" looking for the string "induced" and will return 1 or 2 representing T or F. But how are those two parts interacting in a 'programming' sense? Is there an operator that can be placed in between that might give me some insight?

plot(librarySizeFactors(sce.416b), sizeFactors(sce.416b), pch=16,
    xlab="Library size factors", ylab="Deconvolution factors", 
    col=c("black", "red")[grepl("induced", sce.416b$phenotype)+1],
    log="xy")

When looking at such code, I like to start from the center and work outwards (if that makes sense).

First we have the grepl statement, which looks for matches between the vector sce.416b$phenotype to the string "induced". For each element of sce.416b$phenotype, the grepl function will return a TRUE or a FALSE, i.e., a match or a no-match to "induced".

Second you need to know that R (and many other languages) allow for interpreting TRUE and FALSE as 1 and 0. Try doing TRUE*5 or FALSE-1 in the console, you will see.

So, the +1 behind grepl is converting the 0's and 1's to 1's and 2's, respectively. The result of grepl("induced", sce.416b$phenotype)+1 will thus be a vector of 1's and 2's. Lets say that vector is c(1,1,1,2,2,2).

c("black", "red") is simply a vector of lenght 2. The brackets are just selecting, based on position, either the value at position 1 or the one at position 2. If inside the brackets we have c(1,1,1,2,2,2), then the end result will be c("black","black","black","red","red","red").

As a final note, these are concepts that are not specific to Bioconductor, but rather general to R. So typically, questions like these are raised and answered on platforms like stackoverflow.