Whole genome amplification by multiple displacement amplification (MDA) offers investigators using precious genomic DNA samples a high fidelity method for amplifying nanogram quantities of DNA several thousandfold. This becomes especially important for the modern day genomics researcher who more and more commonly is applying today's genome scanning technologies to patient cohort samples collected years ago that are irrecoverable and invariably in short supply. We present evidence here that MDA-prepared genomic DNA includes artifacts of chromosomal copy number that resemble copy number polymorphisms (CNPs) upon analysis of the DNA on the Affymetrix 10K GeneChip®. The study of CNPs in both health and disease is a rapidly growing area of research, however our current understanding of the relevance of CNPs is incomplete. Our data indicate that utilization of whole genome-amplified samples for analysis heavily reliant on accurate copy number retention could be confounded if the genomic DNA sample was subjected to MDA. We recommend that small amounts of patient cohort DNA stocks be set aside and not subjected to whole genome amplification in order to facilitate the unbiased determination of chromosomal copy numbers when desired.

Introduction

Retaining sufficient amounts of high-quality genomic DNA is often a challenge for researchers performing genetic-based studies. For each recruited patient cohort, larger quantities of genomic DNA are increasingly required for testing due to the rapidly expanding array of high-throughput analyses at the disposal of the modern day genomics researcher. Additionally, genomic DNA samples are often isolated from archived tissue or blood/buccal cell spot collection cards, two sources that primarily yield relatively low amounts of high-quality genomic DNA for further analysis (1,2). Traditional isolation of DNA from freshly purified lymphocytes yields greater amounts of DNA, yet this relatively invasive procedure is not foolproof, sometimes leading to the isolation of lowered amounts of useable DNA, and in most studies is only able to be performed once without the painstaking re-recruitment of the original patient cohort for an additional blood draw that can lead to further experimental design issues (i.e., not all patients are able to be present for a second blood draw). These issues of DNA quality and quantity also confront researchers involved in prenatal, forensic, and anthropology genetic study.

To overcome these issues, several methods for the nonspecific amplification of the entire genome, termed whole genome amplification (WGA), from nanogram quantities of genomic DNA have been developed. PCR-based amplification techniques include degenerate oligonucleotide primed PCR (DOP-PCR) or variations thereof [long DOP-PCR and long, low (LL)-DOP-PCR], primer extension preamplification (PEP), tagged PCR, and balanced PCR amplification. These methodologies all suffer from some sort of limitation, including generation of relatively short fragments (DOP-PCR and PEP), low yield (PEP and tagged PCR), increased technical complexity (tagged PCR), and high levels of input DNA (long DOP-PCR), and there are several publications in the literature detailing modification of these protocols to avoid such limitations (3,4,5,6,7). However, a non-PCR-based method termed multiple displacement amplification (MDA) that uses a φ29 polymerase-mediated reaction carried out in an overnight isothermal reaction, was recently developed (8,9,10). The φ29 polymerase, isolated from bacteriophage φ29, is a desirable enzyme for WGA due to its high processivity at 30°C, strand displacement capacity, 3′→5′ exonuclease activity, and perhaps most importantly, its protein-primed method of replication, which results in the synthesis of DNA fragments >70 kb and error rates as low as 1 mistake in a million (11,12). Studies have shown that MDA produces excellent genotype call concordance between amplified and nonamplified templates (reported error rates of 1 mistake per 9.5 × 10⁻⁶ bp and a 99.95% concordance between MD-WGA and non-MD-WGA genotypes on the Affymetrix 10K GeneChip®), and it has quickly become the preferential WGA technique (8)(13,14,15,16,17).

As the complexity of the human genome and the variations within are further elucidated, it is becoming increasingly evident that small chromosomal copy number alterations or polymorphisms (CNPs) are present across the genome in healthy individuals, however they have been proposed to play a role in disease (18,19,20,21,22). CNPs of various sizes have been identified, however they are commonly <10 kb in size and generally >200 bp (18,19,23). Some investigators have shown that CNPs appear to be in linkage disequilibrium with neighboring single nucleotide polymorphisms (SNPs) in the genome, thereby suggesting that they both share a similar evolutionary history and could also be assayed by proxy through the genotyping of the linked SNP(s) (19). The mechanism of CNP creation includes retrotransposition among other theories that are currently under investigation (20).

Diseases like Down syndrome, trisomy 21, and Prader-Willi syndrome (PWS), a deletion of the q-arm of chromosome 15, are well-documented large chromosomal aberrations that have been suggested to be localized to a much smaller genetic region [the approximate 5-Mb Down syndrome critical region, or DCR, and the small nuclear ribonucleoprotein polypeptide N (SNRPN) locus for a proportion of PWS patients] that can be causative if it alone is altered (24,25). For the analysis of CNPs and disease, it is absolutely critical that if the WGA template is used, there are no artifactual changes in copy number introduced, as such changes would be indistinguishable from any true disease-associated CNPs.