PT  - JOURNAL ARTICLE
AU  - Elizabeth Nacheva
AU  - Katya Mokretar
AU  - Aynur Soenmez
AU  - Alan M Pittman
AU  - Colin Grace
AU  - Roberto Valli
AU  - Ayesha Ejaz
AU  - Selina Vattathil
AU  - Emanuela Maserati
AU  - Henry Houlden
AU  - Jan-Willem Taanman
AU  - Anthony H Schapira
AU  - Christos Proukakis
TI  - DNA isolation protocol effects on nuclear DNA analysis by microarrays, droplet digital PCR, and whole genome sequencing, and on mitochondrial DNA copy number estimation
AID  - 10.1101/151126
DP  - 2017 Jan 01
TA  - bioRxiv
PG  - 151126
4099  - http://biorxiv.org/content/early/2017/06/16/151126.short
4100  - http://biorxiv.org/content/early/2017/06/16/151126.full
AB  - Potential bias introduced during DNA isolation is inadequately explored, although it could have significant impact on downstream analysis. To investigate this in human brain, we isolated DNA from cerebellum and frontal cortex using spin columns under different conditions, and salting-out. We first analysed DNA using array CGH, which revealed a striking wave pattern suggesting primarily GC-rich cerebellar losses, even against matched frontal cortex DNA, with a similar pattern on a SNP array. The aCGH changes varied with the isolation protocol. Droplet digital PCR of two genes also showed protocol-dependent losses. Whole genome sequencing showed GC-dependent variation in coverage with spin column isolation from cerebellum. We also extracted and sequenced DNA from substantia nigra using salting-out and phenol / chloroform. The mtDNA copy number, assessed by reads mapping to the mitochondrial genome, was higher in substantia nigra when using phenol / chloroform. We thus provide evidence for significant method-dependent bias in DNA isolation from human brain, as reported in rat tissues. This may contribute to array “waves”, and could affect copy number determination, particularly if mosaicism is being sought, and sequencing coverage. Variations in isolation protocol may also affect apparent mtDNA abundance.