Strong bias in the location of functional promoter polymorphisms

A considerable proportion of heritable human phenotypic variation is thought to result from altered gene expression. Unfortunately, it is currently impossible to use bioinformatic analysis to discriminate between DNA sequence variants that are likely to influence gene expression and those that are not. In an attempt to define some of the characteristics of promoter polymorphisms with functional effects on gene expression, we examined 674 haplotypes representing 247 unique gene promoters using a standardized reporter gene assay system. Sequence variants that altered gene expression by 1.5‐fold or more were strongly biased toward a location in the core and proximal promoter regions, 50% being within the first 100 bases 5′ to the transcription start site. No bias was seen in the allele frequencies of functional and nonfunctional sequence variants. Only 33% of the functional variants were found in known consensus transcription factor binding sequences or motifs, which suggests that either there are many unknown transcription factor binding motifs or other, unknown mechanisms are involved. The genes with functional polymorphisms that are reported here for the first time include AGTRL2, CAT, CHRNA5, CTSG, CYP2D6, DLD, ERCC1, GABRA1, GABRP, HNRPH3, HIP1, IGKV1‐9, KCNJ15, KCNK6, KLK1, MSMB, MYOC, NPY2R, NOTCH4, ORM2, PEDF, PTPRCAP, ST16 (IL24), SULT1A1, and TSHR. Hum Mutat 26(3), 214–223, 2005. © 2005 Wiley‐Liss, Inc.