Multiplex SNP genotyping in pooled DNA samples by a four-colour microarray system

Abstract

We selected 125 candidate single nucleotide polymorphisms (SNPs) in genes belonging to the human type 1 interferon (IFN) gene family and the genes coding for proteins in the main type 1 IFN signalling pathway by screening databases and by in silico comparison of DNA sequences. Using quantitative analysis of pooled DNA samples by solid-phase mini-sequencing, we found that only 20% of the candidate SNPs were polymorphic in the Finnish and Swedish populations. To allow more effective validation of candidate SNPs, we developed a four-colour microarray-based mini-sequencing assay for multiplex, quantitative allele frequency determination in pooled DNA samples. We used cyclic mini-sequencing reactions with primers carrying 5'-tag sequences, followed by capture of the products on microarrays by hybridisation to complementary tag oligonucleotides. Standard curves prepared from mixtures of known amounts of SNP alleles demonstrate the applicability of the system to quantitative analysis, and showed that for about half of the tested SNPs the limit of detection for the minority allele was below 5%. The microarray-based genotyping system established here is universally applicable for genotyping and quantification of any SNP, and the validated system for SNPs in type 1 IFN-related genes should find many applications in genetic studies of this important immunoregulatory pathway.

Figure 1

(A) Steps of the genotyping procedure. (B) Schematic illustration of the ‘array of arrays’ format. Each microscope slide contains 56 subarrays in a 4 × 14 conformation, and each subarray carries ‘anti-tag’ oligonucleotides in a 10 × 11 spot conformation.

Figure 2

Image at four wavelengths of a row of four ‘subarrays’ on an oligonucleotide array. Four samples have been genotyped for 25 IFN-related SNPs with primers for both DNA strands spotted in duplicate. The respective laser power and PMT gain settings of the array scanner were 90 and 75% for Texas Red-ddATP and R110-ddGTP, 100 and 75% for TAMRA-ddCTP and 100 and 80% for Cy5-ddUTP.

Figure 3

Cluster analysis of the results from typing the SNPs IFNAR-w (A) and STAT1-f (B) in 75 samples using the IFN oligonucleotide array system. The values on the y-axes are logarithms of the sum of background corrected signals for both alleles. The allele ratios on the x-axes are calculated from the background corrected signal intensities for both alleles. The genotype frequencies conform to Hardy–Weinberg equilibrium.

Figure 4

Standard curves for nine SNPs. The SNPs are: (A) IFI27, (B) Jak1-d, (C) IFNAR1-e, (D) IFNAR1-v, (E) IFNAR1-w, (F) IFNAR2-k, (G) IFNAR2-m, (H) STAT1-e, (I) STAT3-14. Samples homozygous for both alleles of each SNP were mixed in appropriate ratios to obtain the allele ratios given on the x-axes. The correlation coefficient (R²) for the curves is given in each panel. The SDs indicated by vertical bars in the curves are calculated from four reactions performed in separate ‘subarrays’. No homozygotes to prepare mixtures of allelic ratios below 0.5 were available for the curves (A) and (H).