Validity of maternal genotypes in DNA from archival pregnancy serum samples.

Much functional genomic research has focused on genotypic data from large cohorts or entire populations (1 ). The large populationbased serum biobanks collected for maternal (pregnancy) screening and stored in many countries could be useful for genetic epidemiologic studies if DNA extracted from archival serum and plasma is of sufficient quantity and quality for successful genotyping analyses (2– 4 ). Serum collected during pregnancy also contains cell-free DNA from the fetus (5 ), however, which might affect the outcome of genotyping analyses. We have evaluated the concordance of genotypes between DNA extracted from archival maternal sera with DNA from fresh whole blood from the same women. The median DNA yield was 15 mg/L (range 1–34 mg/L) using QiaAmp DNA minipreps (Qiagen) on 200 L fresh whole blood from 137 women and 90 g/L (range 0 – 4800 g/L) using the MagNA Pure LC and Total Nucleic Acid Isolation Kit (MagNAPure; Roche Diagnostics) on 200 L of 191 archival maternal sera from the same women. DNA yield decreased with increasing serum sample age (P 0.005, Jonckheere–Terpstra) (Fig. 1), with no difference due to firstor second-trimester sampling (P 0.616, Mann–Whitney). Sample volumes were sufficient for repeat DNA extractions from 175 serum samples. Twenty microliters from each 50L extract was evaporated to dryness at 37 °C and used in a 6L Y chromosome–specific PCR (5 ) with 0.0625 L 40 assay mix, 0.1% BSA (Sigma-Aldrich), and 3.5 mol/L MgCl2 on a 7900 HT Sequence Detection System (Applied Biosystems). Another evaporated 20L aliquot was used for multiple displacement amplification (MDA) (GenomiPhi V2; GE healthcare) with 0.1% BSA. MDA products were dissolved in 80 L TE (10 mmol/L Tris, 1 mmol/L EDTA, pH 8.0) by shaking at 4 °C overnight. DNA concentrations were determined using 2 L template in a F2 g.20210G A (rs1799963) TaqMan MGB assay, with serum samples supplemented with BSA and MgCl2 as above. All samples were analyzed for 10 highfrequency single nucleotide polymorphisms (SNPs) using 2 L whole blood extract, 9 L serum extract, or 2 L MDA product, 0.075 L assay mix, BSA, and MgCl2 as above. Seventy-two of 84 serum samples from women carrying male fetuses and 1 of 91 samples from women carrying female fetuses were Y chromosome positive (P 8E 30). There was no correlation between sample age, gestational age (P 0.2– 0.9, all Pearson ), or DNA yield (P 0.9, Mann–Whitney), and Y chromosome detection. The genotyping rate was 88.9% when using 0.1 ng up to 0.4 ng serum DNA and 99% when using 0.4 ng serum DNA or 2 ng whole blood DNA as PCR template. The DNA yield was enough to provide 0.4 ng per SNP analysis from 81% (154 of 191) of the serum samples. All genotyping results that were based on 0.1 ng archival serum DNA as template were identical to the corresponding whole blood samples. Analysis of an admixture of heterozygous and homozygous DNA for one of the SNPs revealed that at least 50% heterozygous DNA was required to alter the genotyping results of homozygous samples and at least 90% homozygous DNA was required to alter the genotyping results of heterozygous samples. The median DNA yield after MDA was 506 ng (range 0 – 4165 ng). Genotypes of the MDA-treated samples could be automatically determined for 4 of 10 SNPs by the 7900HT software. The genotyping rate was 91%, of which 90% were concordant with corresponding genotypes of neat serum extracts. The poor performance in MDA and subsequent genotyping analyses could be due to inhibitors present in the MagNAPure extracts, DNA degradation and fragmentation during storage, or sequence context. The failed SNPs contained poly G or C runs, CG content 50%, and/or multiple CpG sites. The lowest genotyping and concordance rates were obtained from the MDA products of the serum samples with the lowest DNA yield. Archival serum from maternity serological screening is a useful source of DNA for genetic epidemiologic studies. For reliable genotyping results, at least 0.4 ng of serum-derived DNA should be used in the TaqMan genotyping assays. The presence of realistic amounts of fetal DNA of a discordant genotype may at worst cause an indeterminate genotypic assignment but will not cause false maternal genotyping results on the TaqMan 7900 HT system. MDA 1 Nonstandard abbreviations: MDA, multiple displacement amplification; SNP, single nucleotide polymorphism. Fig. 1. DNA yield from serum in relation to sample age. Clinical Chemistry 55:4 842–843 (2009) Letter to the Editor