I normalized a data using the quantile function in the Bioconductor package limma, boxplots for each group before (A) and after (B) the process this does not look like what I usually see.

Could anyone, please, explain to me why by that black part passed to down of the boxplot in B?

Through quantile normalisation, which I assume you used, it's virtually impossible for the highly-expressed outliers in A to be the same as the lowly-expressed outliers in B. I'm not sure if that's what you were implying?

It may be due to the long, lower tail that's present in one of your samples (9 samples from the right in A), which, through quantile normalisation, would affect the distribution in all other samples.

I neither believe that this is a problem in any way. I have seen elsewhere how, after quantile normalisation, lowly-expressed outliers will exist. Before deciding if this is a problem, you should take a look at:

PCA bi-plot

density plots of normalised counts

pairwise scatterplot matrices

quantile-quantile plots after differential expression analysis

I checked my code and found an error, I used this function to create the boxplot, but my sample x was already in the log function, so it was as if I was log twice.

Hey Kevin,

Thank you so much! I

boxplot(log(as.matrix(

x)),las=2,ylab="Log2(Intensity)")I checked my code and found an error, I used this function to create the boxplot, but my sample

xwas already in the log function, so it was as if I was log twice.Now I gotta fix the error.

Best regards,

Leite