Natural Selection and Functional Potentials of Human Noncoding Elements Revealed by Analysis of Next Generation Sequencing Data

Noncoding DNA sequences (NCS) have attracted much attention recently due to their functional potentials. Here we attempted to reveal the functional roles of noncoding sequences from the point of view of natural selection that typically indicates the functional potentials of certain genomic elements. We analyzed nearly 37 million single nucleotide polymorphisms (SNPs) of Phase I data of the 1000 Genomes Project. We estimated a series of key parameters of population genetics and molecular evolution to characterize sequence variations of the noncoding genome within and between populations, and identified the natural selection footprints in NCS in worldwide human populations. Our results showed that purifying selection is prevalent and there is substantial constraint of variations in NCS, while positive selectionis more likely to be specific to some particular genomic regions and regional populations. Intriguingly, we observed larger fraction of non-conserved NCS variants with lower derived allele frequency in the genome, indicating possible functional gain of non-conserved NCS. Notably, NCS elements are enriched for potentially functional markers such as eQTLs, TF motif, and DNase I footprints in the genome. More interestingly, some NCS variants associated with diseases such as Alzheimer's disease, Type 1 diabetes, and immune-related bowel disorder (IBD) showed signatures of positive selection, although the majority of NCS variants, reported as risk alleles by genome-wide association studies, showed signatures of negative selection. Our analyses provided compelling evidence of natural selection forces on noncoding sequences in the human genome and advanced our understanding of their functional potentials that play important roles in disease etiology and human evolution.非编码DNA序列（Noncoding DNA sequences, NCS）近年来因其功能潜力受到广泛关注。本研究尝试从自然选择的视角揭示非编码序列的功能作用——自然选择通常可反映特定基因组元件的功能潜力。我们分析了千人基因组计划（1000 Genomes Project）第一阶段数据中近3700万个单核苷酸多态性（Single Nucleotide Polymorphisms, SNPs），估算了一系列群体遗传学与分子演化的关键参数，以刻画群体内外非编码基因组的序列变异特征，并在全球人类群体的非编码DNA序列中鉴定出自然选择的印迹。 研究结果显示，净化选择（purifying selection）在非编码DNA序列中普遍存在，非编码基因组的变异受到显著约束；而正选择（positive selection）则更倾向于局限于特定基因组区域与区域性人群。有趣的是，我们观察到基因组中衍生等位基因频率较低的非保守非编码DNA序列变异占比更高，这提示非保守非编码DNA序列可能存在功能获得性。 值得注意的是，非编码DNA序列元件在基因组中显著富集于表达数量性状基因座（expression Quantitative Trait Locus, eQTL）、转录因子（Transcription Factor, TF）结合基序、DNase I足迹等潜在功能标记。更引人关注的是，尽管绝大多数被全基因组关联研究（Genome-Wide Association Studies, GWAS）报道为风险等位基因的非编码DNA序列变异呈现净化选择的特征，但部分与阿尔茨海默病、1型糖尿病、炎症性肠病（Inflammatory Bowel Disease, IBD）等疾病相关的非编码DNA序列变异却表现出正选择的印迹。 本研究为人类基因组中非编码序列受到自然选择压力作用提供了有力证据，并加深了我们对其功能潜力的理解——这些序列在疾病病因学与人类演化中发挥着重要作用。