Comparisons of PCR-based genome amplification systems using CpG island microarrays

The characterization of complex DNA libraries by high-throughput sequencing technologies provides a powerful approach toward finding mutations and genetic variation in the human genome. However, the value of these analyses is highly dependent upon the quality of the DNA libraries themselves. For example, the sequence composition of libraries made using PCR-based procedures can be skewed due to biases in the amplification efficiency of individual library members. Here, we used CpG island microarrays to evaluate the biases incurred in PCR-based genome amplification using three different DNA polymerase mixtures designed to efficiently amplify problematic sequence tracts. Based on hybridization properties of dye-labeled DNA libraries to these microarrays, we quantified the overall and specific trends in the PCR efficiency of more than 1,400 sequences with high GC-content, which generally amplify with low efficiency using conventional PCR protocols. Overall, all three DNA polymerase mixtures produced libraries that show substantial increases in the representation of CpG island segments that poorly amplify with Taq DNA polymerase. However, the effects of these DNA polymerases were quite specific since they did not alter the relative representation of segments that efficiently amplify with Taq DNA polymerase. Furthermore, we demonstrate that DNA microarrays provide a robust platform for rapidly evaluating the ability of different PCR systems to amplify difficult genomic regions targeted for mutational, resequencing, and genotyping analyses.