Unbiased whole-genome amplification directly from clinical samples

Abstract

Preparation of genomic DNA from clinical samples is a bottleneck in genotyping and DNA sequencing analysis and is frequently limited by the amount of specimen available. We use Multiple Displacement Amplification (MDA) to amplify the whole genome 10,000-fold directly from small amounts of whole blood, dried blood, buccal cells, cultured cells, and buffy coats specimens, generating large amounts of DNA for genetic testing. Genomic DNA was evenly amplified with complete coverage and consistent representation of all genes. All 47 loci analyzed from 44 individuals were represented in the amplified DNA at between 0.5- and 3.0-fold of the copy number in the starting genomic DNA template. A high-fidelity DNA polymerase ensures accurate representation of the DNA sequence. The amplified DNA was indistinguishable from the original genomic DNA template in 5 SNP and 10 microsatellite DNA assays on three different clinical sample types for 20 individuals. Amplification of genomic DNA directly from cells is highly reproducible, eliminates the need for DNA template purification, and allows genetic testing from small clinical samples. The low amplification bias of MDA represents a dramatic technical improvement in the ability to amplify a whole genome compared with older, PCR-based methods.

Figure 1.

Comparison of MDA carried out from biological samples. Eight loci were examined for representation by the TaqMan assay. (Buccal swab) Black bars; (whole blood) cross-hatched bars; (finger stick) dotted bars; (buffy coat) gray bars; (Guthrie card) white bar; (control genomic DNA) diagonal-hatched bar. For buccal swab, whole blood, and finger stick blood, the error bar representing 1 SD was generated from the average of 20 different individual MDA amplifications. For buffy coat and Guthrie card, the error bar representing 1 SD was generated from the average of 6 different individual MDA amplifications.

Figure 2.

Amplification bias analysis by the TaqMan assay for 47 human loci. The location on the chromosome for the 47 loci tested from the WIAF (Whitehead Institute-Affymetrix) SNPs database (http://www-genome.wi.mit.edu/snp/human/) is in Table 1. Loci representation relative to the starting DNA template for each of the 47 loci. Each bar represents the average loci representation of 44 patients. The average of all of the 47 loci was 117% (black bar).

Figure 3.

Histograms of loci representations for 47 genetic Loci and 44 DNA samples. Each graph is for a different locus, with each bar representing one of the 44 DNA samples. The loci representation for the Y-chromosome-specific sex determining region gene (SRY) is depicted in the bottom right-hand corner.

Figure 4.

Loci representation for repetitive elements. (A) Long interspersed nuclear element (LINE); (B) short interspersed nuclear element (SINE).

Figure 4.

Loci representation for repetitive elements. (A) Long interspersed nuclear element (LINE); (B) short interspersed nuclear element (SINE).

Figure 5.

Representative GeneScan software electropherograms of ABI AmpFLSTR Profiler Plus STR analysis. MDA-amplified DNA from whole blood (Table 3, sample # 6, WB) was compared with the conventionally purified DNA (Table 3, sample # 6, GD) for genotyping of 9 tetranucleotide short tandem repeats (D3S1358, vWA, FGA, D8S1179, D21S11, D18S51, D5S818, D13S317, D7S820) and the Amelogenin locus.