The utility and limitation of single nucleotide polymorphism analysis on whole genome amplified mesenchymal tumour DNA in formalin fixed tumour samples

Abstract

In the study of tumour genetics, formalin fixed, paraffin embedded (FFPE) tumours are the most readily available tissue samples and DNA derived from FFPE tissue has been validated for array comparative genomic hybridisation (aCGH) and single nucleotide polymorphism (SNP) array analysis. Furthermore, in the study of tumour precursor genetics, whole genome amplification (WGA) has been used to produce a sufficient amount of DNA for aCGH. However, it is unclear whether the same approach can be extended to high-resolution SNP analysis. In this study, we examined the utility and limitations of genotyping platforms performed on WGA DNA from FFPE mesenchymal tumour samples for both copy number and SNP analyses. We analysed the results obtained using DNA derived from matched FFPE and frozen tissue samples on the Affymetrix 250K Nsp SNP array. Two widely used WGA methods, Qiagen (isothermal protocol) and Sigma (thermocycling protocol) were employed to determine how WGA methods affect the results. We found that the use of WGA DNA derived from FFPE mesenchymal tumours for high-resolution SNP array application can produce a significant amount of false positive and false negative findings. While some of these misinterpretations appear to cluster in genomic regions with high or low GC contents, the majority appears to occur randomly. Only large scale chromosome loss of heterozygosity (LOH) (>10 Mb) can be reliably detected from WGA FFPE tumour DNA samples but not smaller LOH or copy number alterations. Our findings here indicate a need for caution in SNP array data interpretation when using WGA FFPE tumour-derived DNA in determining genomic alterations less than 10 Mb.