Search

Functional characterization of disease-causing variants at risk loci has been a significant challenge. Here we report a high-throughput single-nucleotide polymorphisms sequencing (SNPs-seq) technology to simultaneously screen hundreds to thousands of SNPs for their allele-dependent protein-binding differences. This technology takes advantage of higher retention rate of protein-bound DNA oligos in protein purification column to quantitatively sequence these SNP-containing oligos. We apply this technology to test prostate cancer-risk loci and observe differential allelic protein binding in a significant number of selected SNPs. We also test a unique application of self-transcribing active regulatory region sequencing (STARR-seq) in characterizing allele-dependent transcriptional regulation and provide detailed functional analysis at two risk loci (RGS17 and ASCL2). Together, we introduce a powerful high-throughput pipeline for large-scale screening of functional SNPs at disease risk loci., Functional characterization of disease-causing variants at risk loci in cancer is challenging. Here, in prostate cancer the authors report a pipeline for high-throughput single-nucleotide polymorphisms sequencing (SNPs-seq) for large scale screening of functional SNPs at disease risk loci.

Background. By SNP genotyping and RNA sequencing of 471 normal prostate samples, we recently created a prostate tissue-based eQTL dataset and identified significant eQTL signals at 51 prostate cancer risk loci. To functionally characterize these risk SNPs, we developed a massively parallel sequencing technology to screen SNPs for their allele-dependent protein binding differences. We combined this technology (called SNPs-seq) with another high throughput assay (called STARR-seq) to screen the risk loci with significant prostate-specific eQTL signals. Methods. To select candidate functional SNPs in eQTL regions, we took advantage of existing epigenomic datasets and available tools including ENCODE, HaploReg, and Regulome. For all selected SNPs, we first made allele-specific double-strand oligos and performed DNA-protein binding assays. We then performed sequencing analysis on the protein-bound DNA oligos and determined allele-specific protein binding differences. To evaluate reproducibility of SNPs-seq, we performed each assay in duplicates. We cloned SNPs-seq screened SNP regions showing allele-specific protein binding differences into the STARR-seq vector to further determine allele-specific enhancer activities. Finally, we performed EMSA and luciferase reporter assays to validate a set of promising candidate SNPs. Results. From 51 risk loci with strong eQTL signals, we selected 374 SNPs with strong indication of regulatory potential, as evidenced by overlapping with epigenomic marks. When comparing technical duplicates, sequence read counts from the SNPs-seq showed significant correlation with r2>=0.99. By normalizing input controls, we found 101 of the 374 SNPs showing significant allelic protein binding differences (>=1.5-fold binding difference between variant and reference alleles). Interestingly, three published functional SNPs (rs12769019, rs10993994, and rs4907792) were also among the significant SNPs, validating SNPs-seq as functional SNP screening tool. To further validate the candidate SNPs from SNPs-seq, we applied STARR-seq and tested the 101 SNPs-containing sequences (371-686bp) in LNCaP cell line under androgen treatment. This analysis revealed 11 SNPs that not only demonstrated enhancer/repressor activity but also functioned with allelic differences. EMSA and luciferase reporter assays confirmed 6 SNPs with allele-dependent enhancer/repressor activity. Conclusions. We developed a high throughput sequencing-based technology to screen large number candidate SNPs for their allelic protein binding differences. The SNPs-seq coupled with STARR-seq will provide a powerful strategy for functionally characterizing risk loci in prostate cancer and other common diseases. Further understanding genetic role of prostate cancer etiology may facilitate the translation of population-based discovery into biological mechanisms and eventually benefit clinical practice. Citation Format: Peng Zhang, Jing Zhu, Sufyan Suleman, Yong-Chen Guo, Mei-Jun Du, Li-Dong Wang, Gong-Hong Wei, Liang Wang. Functional characterization of prostate cancer risk loci by SNPs-seq and STARR-seq [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1280. doi:10.1158/1538-7445.AM2017-1280