One possible explanation for non-replication is that SNPs within sequences targeted by expression probes may change hybridization efficiency in an allele-specific manner; if that SNP is also correlated with a genotyped variant, false positive eQTLs may result [35]. While 45.3% and 37.2% of Agilent and Illumina probes overlap with a polymorphism found in dbSNP131 or the one thousand genomes project (2010.08.04 release), the frequency distribution of polymorphisms in and around probe sequences differs markedly between the Agilent (UC) and Illumina (UW) platforms (Figure S9); Illumina expression probes have clearly been designed to avoid common polymorphisms.
