Genome-wide association study combined with biological context can reveal more disease-related SNPs altering microRNA target seed sites

Abstract

Background: Emerging studies demonstrate that single nucleotide polymorphisms (SNPs) resided in the microRNA recognition element seed sites (MRESSs) in 3'UTR of mRNAs are putative biomarkers for human diseases and cancers. However, exhaustively experimental validation for the causality of MRESS SNPs is impractical. Therefore bioinformatics have been introduced to predict causal MRESS SNPs. Genome-wide association study (GWAS) provides a way to detect susceptibility of millions of SNPs simultaneously by taking linkage disequilibrium (LD) into account, but the multiple-testing corrections implemented to suppress false positive rate always sacrificed the sensitivity. In our study, we proposed a method to identify candidate causal MRESS SNPs from 12 GWAS datasets without performing multiple-testing corrections. Alternatively, we used biological context to ensure credibility of the selected SNPs.

Results: In 11 out of the 12 GWAS datasets, MRESS SNPs were over-represented in SNPs with p-value ≤ 0.05 (odds ratio (OR) ranged from 1.1 to 2.4). Moreover, host genes of susceptible MRESS SNPs in each of the 11 GWAS dataset shared biological context with reported causal genes. There were 286 MRESS SNPs identified by our method, while only 13 SNPs were identified by multiple-testing corrections with a given threshold of 1 × 10-5, which is a common cutoff used in GWAS. 27 out of the 286 candidate SNPs have been reported to be deleterious while only 2 out of 13 multiple-testing corrected SNPs were documented in PubMed. MicroRNA-mRNA interactions affected by the 286 candidate SNPs were likely to present negatively correlated expression. These SNPs introduced greater alternation of binding free energy than other MRESS SNPs, especially when grouping by haplotypes (4210 vs. 4105 cal/mol by mean, 9781 vs. 8521 cal/mol by mean, respectively).

Conclusions: MRESS SNPs are promising disease biomarkers in multiple GWAS datasets. The method of integrating GWAS p-value and biological context is stable and effective for selecting candidate causal MRESS SNPs, it reduces the loss of sensitivity compared to multiple-testing corrections. The 286 candidate causal MRESS SNPs provide researchers a credible source to initialize their design of experimental validations in the future.

Figure 1

Comparisons of the SNP density between different mRNA regions grouped by global MAF. SNPs with global MAF ≥ 0.01 are defined as common SNPs, the rest are rare SNPs. The figure illustrates that the MRESS region has a significantly lower common SNP density but almost the same rare SNP density compared to the UTR regions, which indicates the MRESS region is under strong purifying selection. (Of note, the exceptionally high rare SNP density of the CDS region is derived from its unparalleled sequencing coverage).

Figure 2

Schematic diagram for the construction of the microRNA-gene-disease three-way interactions. Ways to build up relationships between any 2 of the 3 elements are marked above the double-headed arrows. If a microRNA and a gene share the same related disease, we believe that their interaction is also associated with the disease, and any factor such as MRESS SNP that can interfere with this interaction is highly susceptible.

Figure 3

Quantile-Quantile plot for MRESS SNPs in all the 12 GWAS datasets. The units for both x-axis and y-axis are –log10(p). We can observe upward deviation from the red line in most of the subplot, which is the clue for true positive signals, and imply MRESS SNPs as common factors involved in multiple diseases and cancers.

Figure 4

Workflow of our method for selecting candidate causal MRESS SNPs from multiple GWAS datasets. We integrated functional annotations from GO database to filter out false positive MRESS SNPs rather than to perform multiple-correction. The obtained causal SNPs are greatly increased. We further validated our results by three different approaches. The results proved our method stable and effective.

Figure 5

Enrichment pattern of Pearson correlation coefficients of the co-expressed microRNA-mRNA pairs containing candidate causal MRESS SNPs. We divided pairs with candidate causal MRESS SNPs into ten aliquots according to the correlation coefficients and obtained ten intervals. Then, we counted the number of MRESS SNPs, not necessarily candidate SNPs, falling into each interval. At last, we divided the number of candidate causal MRESS SNPs in each interval by that of the total set and took logarithm. A value greater than zero means microRNA-mRNA pairs containing candidate causal MRESS SNPs enrich in this part, and a value less than zero stands for anti-enrichment. This figure displays significant enrichment at parts of weakly negative correlation and anti-enrichment at parts of positive correlation, which demonstrates pairs containing candidate causal MRESS SNPs are more likely to be interactive.

Figure 6

Distribution of the alternations of binding free energy (ΔΔG) are different between candidate causal MRESS SNPs and the total set of MRESS SNPs in 12 GWAS datasets. Red dotted line denotes the candidate causal MRESS SNPs while black solid line denotes the total set of MRESS SNPs in 12 GWAS datasets. ΔΔG in Figure 6 A is calculated by single SNP while that in Figure 6 B is calculated by haplotype. We can observe a decline of the peak near zero and rightward shift for candidate MRESS SNPs, which indicates the greater alternation of binding free energy are created.