I have some microarray dataset like this http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE1935. Given this gene expression data, I want to find out which protein function is likely to be predicted by this gene expression data. I mean genes are related to proteins, So given this gene expression data, I want to know which protein function is likely to be predicted by this data.
if you mean protein "expression," there is probably a positive correlation between mRNA and protein levels but it is hard to say with certainly that any given protein will change in abundance if there is a change in expression.
I don't think gene expression is generally a good way to predict protein "function". The only exception I can think of is regulator genes (such as transcription factors) that affect downstream genes that you may be able to predict using the master regulator function in IPA (if your institute has a license), GSEA, BD-Func, etc.
