[BioC] boxplots: possible unexpected behavior

[Ramon Diaz-Uriarte]

Dear All,

Just a note to alert others who might be getting unexpected results when using 
boxplot.

If you type something like:

boxplot(data$M ~ col(data$M))

any row (gene or spot) that has one or more missing values is not shown in the 
boxplot. In other words, the boxplots will only show the data for those rows 
(those genes or spots) that have no missing values over all the arrays. 


You do not get this behavior (i.e., you get all the points with data to be 
represented for each array) if you either:

a) use a data frame:

boxplot(as.data.frame(data$M))

b) set "options(na.action = "na.pass")"


See my (unjustified) bug-report and B.D. Ripley's answer at 
http://r-bugs.biostat.ku.dk/cgi-bin/R/incoming?id=6846;user=guest;selectid=6846


Best,

R.

-- 
Ramón Díaz-Uriarte
Bioinformatics Unit
Centro Nacional de Investigaciones Oncológicas (CNIO)
(Spanish National Cancer Center)
Melchor Fernández Almagro, 3
28029 Madrid (Spain)
Fax: +-34-91-224-6972
Phone: +-34-91-224-6900

http://bioinfo.cnio.es/~rdiaz
PGP KeyID: 0xE89B3462
(http://bioinfo.cnio.es/~rdiaz/0xE89B3462.asc)