The OS is not inmportant when R is used. I reccomend that you use Rstudio or Rcommander for program the analysis. You can install Rcommander in R console with the command

install.package("Rcommander")

Rcommander has a GUI for avoiding programming. But if you want to take R adventages you must learn to program in R and for it is better Rstudio. Rstudio provide a interface to program very similar to matlab GUI. Both can be used in linux systems, Windows or Mac.

http://www.rcommander.com/

http://www.rstudio.com/

# This example can run in your computer perfectly

# We check the number of cylinders effect in car gasoline consumption( miles per gallon )

d F)

# cyl 2 398891 199445 74.83 3.55e-12 ***

# Residuals 29 77294 2665

# ---

# Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

# > TukeyHSD(aov(formula= disp ~ cyl, data= d))

# Tukey multiple comparisons of means

# 95% family-wise confidence level

#

# Fit: aov(formula = disp ~ cyl, data = d)

#

# $cyl

# diff lwr upr p adj

# 6-4 78.17792 16.53272 139.8231 0.0106851

# 8-4 247.96364 196.59264 299.3346 0.0000000

# 8-6 169.78571 110.76493 228.8065 0.0000002

As Javier mentioned above, the OS is not important if you're running functions that are already in R or can be installed via some R repository such as Bioconductor or CRAN.

Speaking for myself, I only care about the OS if I want to run any system commands within my R scripts.

With linux, you'll have all the goodies of the command line which make things a lot easier e.g. for extracting stuff out of files you have grep and awk:

Grep.Lines