To DNAse or Not DNAse?
We, like most microarray users, have always believed that it was essential to remove any contaminating genomic DNA from our RNA preps before working them up for running on an array. To see if this was completely necessary, we prepared RNA from a mouse brain sample using the Autogen 245T or Trizol methods. No genomic DNA was apparent on a BioAnalyzer tracing with either method. After isolation, each sample was then split and one half repurified using a Qiagen RNAeasy column (Fig.1). The four resulting samples were then run on Affymetrix U430 2.0 arrays. The data was normalized with RMA and all pairwise comparisons performed. To our surprise, there was no detectable difference before and after DNAse treatment (Fig. 2). We subsequently repeated, and confirmed, these results using a mouse liver sample.
Based on these results, we now recommend that DNAse treatment may not be necessary if there is no detectable genomic DNA presence on the BioAnalyzer and you are using a microarray with multiple probes per transcript (the more the better). We believe that while genomic DNA can give aberrant signal, arrays with multiple probes tend to discard these signals as probe set outliers and thus don’t show up in the final analysis. We believe DNAse treatment is still required for any array that uses a few or single probes per transcript.
