Hello all,
I have a matrix of gene-expression data from 50 samples whose rows are the probes and whose columns are the samples. There is an additional column consisting of the Entrez gene IDs for the probes. I also have a vector containing a list of all the genes (Entrez IDs) that this matrix shares with one from another experiment.
How would I subset this matrix to select only the probes (rows) whose Entrez ID is in the vector?
I have tried this code already: myMatrix[myMatrix[,"ID"] %in% commonGenes,]
where myMatrix is my data matrix, ID is the name of the column containing the Entrez IDs, and commonGenes is the character vector containing the list of all the desired Entrez IDs; but this doesn't work (I still get back all the rows.)
you have a matrix or a data.frame? str(myMatriz). May you provide a sample of your data and the error message?
Have you also checked what myMatrix[,"ID"] %in% commonGenes returns?
Paulo and Olena, I don't get an error message, but R just returns all the rows without subsetting. To check, I used nrow(myMatrix) and nrow(myMatrix[myMatrix[,"ID"] %in% commonGenes,]), and they both return the same number.
Here are some screenshots of some commands I entered and their resulting output. The last two screenshots are part of the same output, I just had to scroll down to get it all.
(GSE31547.rma.expressionvalues is the name of the matrix, and GeneSymbol is the name of the column in the matrix whose values I want to select according to the vector common.genes.)
http://postimg.org/gallery/hsjwyicc/
my opinion is that your commonGenes may contain all the Entrez IDs
have a look at this short example (the 6th column playing the role of your Entrez IDs)
M
Hi Long
In principle if a condition applied to the rows of a data.frame does not reduce them to a subset, then this condition is TRUE for all of them. Your condition is (GSE31547.rma.expressionvalues$GeneSymbol %in% common.genes) an it does not reduce your original set, then all of your GeneSymbol are in common.genes.
Following the description of your problem I think you should use the filtering condition GSE31547.rma.expressionvalues$EntrezID %in% common.genes
If this continues failing to subsetting your original data, the first comment (the same posted by Jean Maccario) is valid.
Jorge Ortiz
Jean and Jorge, thank you for your comments. I went back and checked the expression data itself and I found out that I had duplicate probes. Removing the duplicates gave me the expected number of rows when subsetting. Thanks again.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software