(A) Sequence prototypes of the oligonucleotide probes. Twenty-five-mer oligonucleotides which are complementary to SNP sites and flanking sequences are synthesized on the surface of the array. The 13th nucleotide is the interrogative position where the probe sequences are either perfectly matched (PM) or mismatched (MM) to one of the two alleles of the SNP. The PM and MM probe pairs provide a basis for signal vs. noise measurements. The two probe pairs corresponding to the two alleles are grouped as probe quartets. Shown are the prototype sequences of the probe quartet at the SNP site, where the probe sequences differ only at the SNP site which is also the interrogative position. To provide data redundancy, four additional probe quartets are offset from the SNP site by one to four nucleotides in either direction. Also shown are prototype sequences for the probe quartet offset by –4. In this offset probe quartet, the SNP site has shifted to position 17 of the 25-mer. The probe sequences in this quartet are different at –4 (PM vs. MM) and at the SNP site (allele A vs. B). Each SNP is represented by five probe quartets (one at the SNP site and four offset) in both orientations, for a total of 40 oligonucleotide probes. (B) Genotype calling. The RAS is a measure of the signal intensities contributed from the A allele compared to signals from both alleles. Median RAS values from forward and reverse probe quartets define points in x–y coordinates. RAS points from the algorithm training set of 133 ethnically diverse individuals were clustered to determine the three median points, shown as large gray points. Genotypes are assigned for each SNP by determining the shortest distance between RAS points and one of the three median points. Call zones are drawn around the median points to increase the stringency of the genotype assignments. Shown are the RAS points from 99 individuals in 33 CEPH trios for TSC SNP 44146. One of the individuals was not assigned a genotype because the RAS point fell just outside the AB call zone.
