WGCNA analysis of gene expression data results in sevral genes modules labelled with colors and/or numerics. The module labeled grey as a color or as  0 in numerics correponds to the set of genes which have not been clustered in any module. My precise question is that is it there any limit on its size like it should not exceed the 20%, or 30% or 50% of the total data? 

My data has shown that grey module is almost 50% of the total data. My data is from heterogeneous samples and it does make sense to me but not to reviewers. 

There is no limit on the size of any module or the number of the grey genes from the point of view of WGCNA itself. How many genes should end up grey depends a lot on how many genes are in the analysis to begin with and how they were filtered, how selective the analyst wants to be in the module definition - some people prefer tighter modules with only highly correlated genes, others prefer to assign as many genes as possible to modules. I don't think there is a "right" or "wrong" number of grey genes, at least not without knowing much more about how the analysis was carried out.