Human polymorphism at microRNAs and microRNA target sites

Abstract

MicroRNAs (miRNAs) function as endogenous translational repressors of protein-coding genes in animals by binding to target sites in the 3' UTRs of mRNAs. Because a single nucleotide change in the sequence of a target site can affect miRNA regulation, naturally occurring SNPs in target sites are candidates for functional variation that may be of interest for biomedical applications and evolutionary studies. However, little is known to date about variation among humans at miRNAs and their target sites. In this study, we analyzed publicly available SNP data in context with miRNAs and their target sites throughout the human genome, and we found a relatively low level of variation in functional regions of miRNAs, but an appreciable level of variation at target sites. Approximately 400 SNPs were found at experimentally verified target sites or predicted target sites that are otherwise evolutionarily conserved across mammals. Moreover, approximately 250 SNPs potentially create novel target sites for miRNAs in humans. If some variants have functional effects, they might confer phenotypic differences among humans. Although the majority of these SNPs appear to be evolving under neutrality, interestingly, some of these SNPs are found at relatively high population frequencies even in experimentally verified targets, and a few variants are associated with atypically long-range haplotypes that may have been subject to recent positive selection.

Fig. 1.

SNPs in human pre-miRNAs. (A) SNP density in validated human pre-miRNAs and flanking regions. Flanking regions 1–3 represent successive, nonoverlapping windows of ≈100 nt (equal to the size of the given pre-miRNA) located immediately adjacent to the pre-miRNA. Error bars represent the standard deviation of the mean value. (B) Distribution of human SNPs across the secondary structure of pre-miRNAs. The pre-miRNA is apportioned in the following regions: seed (black), MIR^Δseed (light shaded), MIR*, stem (neither MIR or MIR*), and loop. The number of SNPs identified in dbSNP for each region in human pre-miRNAs is denoted for each region.

Fig. 2.

SNP density within computationally predicted target sites and flanking regions. Flanking regions 1–3 represent successive, nonoverlapping windows of 7–8 nt (equal to the size of the given target) located immediately adjacent to the target. Error bars represent the standard deviation of the mean value.

Fig. 3.

Frequencies of minor alleles for SNPs in computationally predicted target sites. SNPs with population frequency estimates from any population in dbSNP were categorized into four categories based on minor allele frequency information: (i) SNPs without frequency information available for any population (category n/a), (ii) SNPs that are monomorphic in all populations genotyped (i.e., minor allele frequency q = 0) (category 0), (iii) SNPs that have been genotyped in at least one population and show a maximal frequency of 0 < q < 0.10 in the population(s) surveyed (category 1), and (iv) SNPs that have been genotyped in at least one population and show a frequency of q ≥ 0.10 in at least one population (category 2). Because a given SNP may have different frequencies in different populations, a SNP was categorized based on its highest frequency in any given population.

Fig. 4.

Distribution of iHS estimates for SNPs in miRNA target sites. ∣iHS∣ values (18) were obtained for category 2 SNPs in computationally predicted sites (black triangles) and in NBR sites (gray circles) for each of the three HapMap populations (i.e., YRI, CEU, or ASN). Note that iHS values were not available for some SNPs, including those found in experimentally verified targets.

Fig. 5.

EHH at various genetic distances for core haplotypes around rs1042538 in the YRI panel. The core haplotype bearing the derived allele at rs1042538 is marked in bold. All other haplotypes from the defined core are marked in gray and hatched lines. The population frequency of each core haplotype is denoted in box.