Let the input file where you have the data be called input.txt.

We will call the first column containing the marker information x and the second column containing the test statistic y. This file can include in its top row the characters x y. This is called a header. The choice is up to the user to put the header in or not.

For entering R on the Linux machine on prompt type:


Once you are in R the steps for using BARS are as follows:

1) Load the shared object library:

    dyn.load("barsN.so",now = F)

2) Source the R wrapper file:


3) Read in the data file,input.txt:

     With header:

            l = read.table("input.txt",header = T)



    Without header:


             l  =  read.table("input.txt")

             x =  l[,1]

             y =  l[,2]


4) Run BARS


    out  =  barsN.fun(x,y)


5) To view all of the BARS results (NOTE: the output may be quite lengthy, so you may want to access particular portions of the output using the $ operator - see below.):




There are certain other commands which we can use for viewing the results:


1) To plot the x and y values.


        plot( x ,y , xlab = 'x values', ylab = 'yvalues')


2) To fit a curve using BARS data


        lines(x , out$postmodes)


3) To find the maximum height of the BARS curve




4) To find the location of the maximum height of the BARS curve




5)  To find the confidence interval for the peak location




    The BARS test is significant if this interval is a proper subset of the range defined

    by min(x), max(x). to see what this interval is,





6) To find the distribution of the peak locations




7) To change the level of confidence in the test to something other than 95% say 99%,

    use the option "conf = 0.99" in the barsN.fun command



           out  =  barsN.fun(x,y,conf = 0.99)


Note :  Another Method to fit a curve using the smooth.spline function


                lines(smooth.spline( x, y, cv = T))


            To find the maximum height of the smooth.spline curve


                    ssout = smooth.spline( x , y , cv = T)

                    lm = predict(ssout)





To find more details of the parameters and options available for BARSN please

have a look at BARSN_options.