A comparison of methods for DNA preparation prior to microarray analysis.

Microarrays are a valuable tool for analysis of both bacterial and eukaryotic nucleic acids. As many of these applications use non-specific amplification to increase sample concentration prior to analysis, the methods used to fragment and label large amplicons are important to achieve the desired analytical selectivity and specificity. Here, we used eight sequenced ESKAPE pathogens to determine the effect of two methods of whole genome amplicon fragmentation and three methods of subsequent labeling on microarray performance; nick translation was also assessed. End labeling of both initial DNase I-treated and sonication-fragmented amplicons failed to provide detectable material for a significant number of sequence-confirmed genes. However, processing of amplicons by nick translation, or by sequential fragmentation and labeling by Universal Labeling System or Klenow fragment/random primer provided good sensitivity and selectivity, with marginally better results obtained by Klenow fragment labeling. Nick-translation provided 91-100% sensitivity and 100% specificity in the tested strains, requiring half as many manipulations and less than 4h to process samples for hybridization; full sample processing from whole genome amplification to final data analysis could be performed in less than 10h. The method of template denaturation before amplification did affect detection sensitivity/selectivity of nick-labeled amplicons, however.