Evaluation of Target Preparation Methods for Single-Feature Polymorphism Detection in Large Complex Plant Genomes

For those genomes low in repetitive DNA, hybridizing total genomic DNA to high-density expression arrays offers an effective strategy for scoring single-feature polymorphisms (SFPs). Of the 2.5 gigabases that constitute the maize (Zea mays L.) genome, only 10 to 20% are genic sequences, with large amounts of repetitive DNA intermixed throughout. Therefore, a target preparation method engineered to generate a high genic-to-repetitive DNA ratio is essential for SFP detection in maize. To that end, we tested four gene enrichment and complexity reduction target preparation methods for scoring SFPs on the Affymetrix GeneChip Maize Genome Array ("Maize GeneChip"). Methylation filtration (MF), C(o)t filtration (CF), mRNA-derived cRNA, and amplified fragment length polymorphism (AFLP) methods were applied to three diverse maize inbred lines (B73, Mo17, and CML69) with three replications per line (36 Maize GeneChips). Our results indicate that these particular target preparation methods offer only modest power to detect SFPs with the Maize GeneChip. Most notably, CF and MF are comparable in power, detecting more than 10 000 SFPs at a 20% false discovery rate. Although reducing sample complexity to approximately 125 megabase by AFLP improves SFP scoring accuracy over other methods, only a minimal number of SFPs are still detected. Our findings of residual repetitive DNA in labeled targets and other experimental errors call for improved gene-enrichment methods and custom array designs to more accurately array genotype large, complex crop genomes.