A LECAMWASAM1,2,3, J CARMODY1,2, E EKINCI2, K DWYER3, R SAFFERY1,4
1Clinical and Disease Epigenetics Group, Murdoch Childrens Research Institute, Victoria, Australia; 2Department of Endocrinology, Austin Health, Victoria, Australia; 3School of Medicine, Faculty of Health Deakin University, Victoria, Australia; 4Department of paediatrics, University of Melbourne, Victoria, Australia
Aim: To characterise the genomic DNA yield from urine and quality of derived methylation data generated from the widely used Illuminia Infinium MethylationEPIC (HM850K) platform and compare this with buffy coat samples.
Background: DNA methylation (DNAm) is the most widely studied epigenetic mark and variations in DNAm profile have been implicated in diabetes which affects approximately 1.7 million Australians.
Methods: QIAamp Viral RNA Mini Kit and QIAamp DNA micro kit were used to extract DNA from frozen and fresh urine samples as well as increasing volumes of fresh urine. Matched buffy coats (BC) to the frozen urine were also obtained and DNA was extracted from the BC using the QIAamp DNA Mini Kit. Genomic DNA (gDNA) of greater concentration than 20ug/ml were used for methylation analysis using the HM850K array.
Results: Irrespective of extraction technique or the use of fresh vs frozen urine samples, limited gDNA was obtained using a starting sample volume of 5ml (0-0.86ug/mL). In order to optimize the yield, we increased starting volumes to 50ml fresh urine, which yielded only 0-9.66ug/mL. A different kit, QIAamp DNA Micro Kit, was trialled in six fresh urine samples and ten frozen urine samples with inadequate DNA yields from 0-17.7ug/mL and 0-1.6ug/mL respectively. Sufficient gDNA was obtained from only 4 of the initial 41 frozen urine samples (10%) for DNA methylation profiling. In comparison, all four BC samples (100%) provided sufficient gDNA.
Conclusion: High quality data can be obtained provided a sufficient yield of gDNA is isolated. Despite optimizing various extraction methodologies, the modest amount of gDNA derived from urine, may limit the generalisability of this approach for the identification of DNAm biomarkers of kidney disease.