Nested if else statements over a number of columns

Question

I have a large data.frame where the first three columns contain information about a marker. The remaining columns are of numeric type for that

Answer 1

These is an other possible solution. All solution above are valid.

My solution is create a function for your case-sensitive without the use of a new library. It's quite long and it's possible to compact, but it's useful to see each step in order to understand how the function works.

olddf <- data.frame(marker = c("kgp5209280_chr3_21902067",
        "chr3_21902130_21902131_A_T",
        "chr3_21902134_21902135_T_C"),
        alleleA = c("T","A","T"),
        alleleB = c("A","T","C"),
        X818 = c(0.0000,0.8626,0.6982),
        X818.1 = c(1.0000,0.1356,0.2854),
        X818.2 = c(0.0000,0.0018,0.0164),
        X345 = c(1.0000,0.7676, 0.5617),
        X345.1 = c(0.0000, 0.2170, 0.3749),
        X345.2 = c(0.0000, 0.0154, 0.0634),   
        X346 = c(0.0000, 0.8626, 0.6982),
        X346.1 = c(1.0000,0.1356, 0.2854), 
        X346.2 = c(0.0000, 0.0018, 0.0164))


mergeallele <- function(arguments,threshold = 0.8){
    n <- nrow(arguments)
    # Creation of a results object as an empty list of length NROW
    # speed for huge data.frame 
    new.lst <- vector(mode="list", n)
    for (i in 1:n){
        marker_row <- arguments[i,]
        colvalue.4 <- NaN
        if (max(marker_row[,c(4:6)]) < threshold){
            colvalue.4 <- max(marker_row[,c(4:6)])
        }

        colvalue.5 <- NaN       
        if (max(marker_row[,c(7:9)]) < threshold){
            colvalue.5 <- max(marker_row[,c(7:9)])
        }

        colvalue.6 <- NaN       
        if (max(marker_row[,c(10:12)]) < threshold){
            colvalue.6 <- max(marker_row[,c(10:12)])
        }
        new.lst[[i]]  <- data.frame(marker_row[,1],
            marker_row[,2],
            marker_row[,3],
            colvalue.4,
            colvalue.5,
            colvalue.6)     
    }   
    new.df <- as.data.frame(do.call("rbind",new.lst))
    names(new.df) <-  c(colnames(arguments)[1],
                    colnames(arguments)[2],
                    colnames(arguments)[3],
                    colnames(arguments)[4],
                    colnames(arguments)[7],
                    colnames(arguments)[10])
    return(new.df)
}


newdf <- mergeallele(olddf)

                      marker alleleA alleleB   X818   X345   X346
1   kgp5209280_chr3_21902067       T       A    NaN    NaN    NaN
2 chr3_21902130_21902131_A_T       A       T    NaN 0.7676    NaN
3 chr3_21902134_21902135_T_C       T       C 0.6982 0.5617 0.6982

about:

threshold = 0.8 

you can set your the threshold value (ex: 0.8) avoid to change variable inside the function

new.lst <- vector(mode="list", n)

you can create a empty list of length the old data.frame and the elements of the list are then gradually filled with the loop results (much faster). See the test speed from this Blog