If you need it in Excel, why don't you do it in Excel?

If R, you get the row means with rowMeans(). To get the variances you will have to apply() the function var() to the rows. Here is an example code, assuming that the data is in a 54675x17 data.frame or matrix "df":

rm

If you need it in Excel, why don't you do it in Excel?

If R, you get the row means with rowMeans(). To get the variances you will have to apply() the function var() to the rows. Here is an example code, assuming that the data is in a 54675x17 data.frame or matrix "df":

rm

Hallo Hasan,

This is not a big deal in R you can do this with the base functions.

### read your data

df

Dr. Wilhelm, thank you a lot for your attention and detailed explanation on the matter.

Hi Gianluca,

I'm new in R programming language and I was not know the matrixStats package. Thank you for your advice.

Hello Stefan,

You are right, it is not a big deal but somehow I could not do it but thank you for your simplified explanation.

Jochen Wilhelm and Stefan Krottenthaler Your answer is very detailed. Thank you. You have added a lot of useful information about the program. Thank you for your cooperation.

Dear Hasan,

You can use a function with a loop to do that:

rmeans

Dr. Mohamed, thank you for your detailed explanation and effort.

In my opinion R is best to deal with row or column means, variances or any other function. First convert your data into csv file, upload it in R and use apply() function. You can also save the results in matrix or vectors form.

regards

Dr Farah, I will try your suggestion. Thank you for your response.

Best,

# on my system, apply is also really slow for huge matrices.

# matrixStats::rowVars is about 100x faster:

> system.time(apply(X,1,var,na.rm=TRUE))

user system elapsed

19.61 0.15 19.91 > system.time(rowVars(X))

user system elapsed

0.2 0.0 0.2

# but one can spare some time by manually calculating the variance

> system.time({

+ n

How complicated can such simple tasks be made?

@Ahmed, what if df contains NA values?

@Ahmed,

the denominator is not right. The number of NA values must be substracted from (dim(.)[2]-1).

Dr @JochenWilhem, I have tried the way that you suggested before for calculation of row - wise mean and variance but there are many NA values in my data frame. So, I used to "Absent/Present" calls from the Affymetrix algorithm to flag genes with questionable expression levels. What I wonder that how I can easily get rid of from questionable expression levels (NA's). What is your recommendation? I have also attached the A/P calls file.

Thanks,

Given X is the numeric matrix, you can subset the rows to those containing at least m finite values (not NA) with this code:

s

Dr. @JochenWilhem , it gives an error in second line, says that object 'm' not found.

@AhmedElwan , I tried your suggestion and it gives an error like, Error in rowMeans(x, na.rm = T) : 'x' must be numeric'. I have attached the CSV file that I am working on.

@Hasan: 'm' is a placeholder for the minimum number of finite value you want.

@JochenWilhelm, Now I see. Thank you very much for your interest Dr. Jochen.