NeuroX, a fast and efficient genotyping platform for investigation of neurodegenerative diseases

Abstract

Our objective was to design a genotyping platform that would allow rapid genetic characterization of samples in the context of genetic mutations and risk factors associated with common neurodegenerative diseases. The platform needed to be relatively affordable, rapid to deploy, and use a common and accessible technology. Central to this project, we wanted to make the content of the platform open to any investigator without restriction. In designing this array we prioritized a number of types of genetic variability for inclusion, such as known risk alleles, disease-causing mutations, putative risk alleles, and other functionally important variants. The array was primarily designed to allow rapid screening of samples for disease-causing mutations and large population studies of risk factors. Notably, an explicit aim was to make this array widely available to facilitate data sharing across and within diseases. The resulting array, NeuroX, is a remarkably cost and time effective solution for high-quality genotyping. NeuroX comprises a backbone of standard Illumina exome content of approximately 240,000 variants, and over 24,000 custom content variants focusing on neurologic diseases. Data are generated at approximately $50-$60 per sample using a 12-sample format chip and regular Infinium infrastructure; thus, genotyping is rapid and accessible to many investigators. Here, we describe the design of NeuroX, discuss the utility of NeuroX in the analyses of rare and common risk variants, and present quality control metrics and a brief primer for the analysis of NeuroX derived data.

Keywords: Genetics; Genotyping; Imputation; Meta-analysis; Methods; Neurodegeneration; Parkinson's.

Figure 1

For both NeuroX custom content and the standard content included on the array, a majority of SNPs across the minor allele frequency spectrum have GenTrain scores > 0.7, suggesting quality genotype clusters are readily available. Discrepancies across content type are partially due to genotype cluster method differences between the two sets of variants (custom and standard content), but also due to the inclusion of rare and difficult to genotype loci in the custom content of the array.

Figure 2

Autosomal variant coverage per mega-base for different content classes. Panel A, custom content coverage; Panel B, standard content coverage; Panel C, coverage for successfully imputed variants (imputation quality > 0.30).

Figure 2

Autosomal variant coverage per mega-base for different content classes. Panel A, custom content coverage; Panel B, standard content coverage; Panel C, coverage for successfully imputed variants (imputation quality > 0.30).

Figure 2

Autosomal variant coverage per mega-base for different content classes. Panel A, custom content coverage; Panel B, standard content coverage; Panel C, coverage for successfully imputed variants (imputation quality > 0.30).