hello every one,
I have a question about the difference between the design co-expression network with RNA-SEQ data and microarray data?
I know that the rna-seq method is new but the microarray method is the oldest, but don't know about other difference.
There should not be any major difference. What you need to consider is the distribution of the data that is used as input to construct the co-expression network. Typically, data that follows a normal distribution (i.e., parametric) would be expected. If your microarray data derives from an Affymetrix array and was processed via the RMA method, then it should already be fine. If RNA-seq, you should ideally normalise the raw counts and transform these via, e.g., rlog or vst (DESeq2), or log CPM (EdgeR).
Check the data's distribution via hist(), and also verify the instructions for the co-expression network construction algorithm / tool / program that you are using.
Most co-expression networks are constructed by first deriving a correlation matrix. In this case, you have to consider the differences between Spearman correlation (non-parametric) and Pearson correlation (parametric).
thank you
can we detect microRNA and lncRNA in two methods with a co-expression network?
I have another question, can I mix RNAseq data and microarray data to design a co-expression network?