Yamaguchi-Kabata Yumi, Shimada Makoto K, Hayakawa Yosuke, Minoshima Shinsei, Chakraborty Ranajit, Gojobori Takashi, Imanishi Tadashi
Biological Information Research Center, National Institute of Advanced Industrial Science and Technology, Tokyo, Japan.
PLoS One. 2008;3(10):e3393. doi: 10.1371/journal.pone.0003393. Epub 2008 Oct 14.
A great amount of data has been accumulated on genetic variations in the human genome, but we still do not know much about how the genetic variations affect gene function. In particular, little is known about the distribution of nonsense polymorphisms in human genes despite their drastic effects on gene products.
METHODOLOGY/PRINCIPAL FINDINGS: To detect polymorphisms affecting gene function, we analyzed all publicly available polymorphisms in a database for single nucleotide polymorphisms (dbSNP build 125) located in the exons of 36,712 known and predicted protein-coding genes that were defined in an annotation project of all human genes and transcripts (H-InvDB ver3.8). We found a total of 252,555 single nucleotide polymorphisms (SNPs) and 8,479 insertion and deletions in the representative transcripts in these genes. The SNPs located in ORFs include 40,484 synonymous and 53,754 nonsynonymous SNPs, and 1,258 SNPs that were predicted to be nonsense SNPs or read-through SNPs. We estimated the density of nonsense SNPs to be 0.85x10(-3) per site, which is lower than that of nonsynonymous SNPs (2.1x10(-3) per site). On average, nonsense SNPs were located 250 codons upstream of the original termination codon, with the substitution occurring most frequently at the first codon position. Of the nonsense SNPs, 581 were predicted to cause nonsense-mediated decay (NMD) of transcripts that would prevent translation. We found that nonsense SNPs causing NMD were more common in genes involving kinase activity and transport. The remaining 602 nonsense SNPs are predicted to produce truncated polypeptides, with an average truncation of 75 amino acids. In addition, 110 read-through SNPs at termination codons were detected.
CONCLUSION/SIGNIFICANCE: Our comprehensive exploration of nonsense polymorphisms showed that nonsense SNPs exist at a lower density than nonsynonymous SNPs, suggesting that nonsense mutations have more severe effects than amino acid changes. The correspondence of nonsense SNPs to known pathological variants suggests that phenotypic effects of nonsense SNPs have been reported for only a small fraction of nonsense SNPs, and that nonsense SNPs causing NMD are more likely to be involved in phenotypic variations. These nonsense SNPs may include pathological variants that have not yet been reported. These data are available from Transcript View of H-InvDB and VarySysDB (http://h-invitational.jp/varygene/).
人类基因组中的遗传变异已积累了大量数据，但我们对这些遗传变异如何影响基因功能仍知之甚少。特别是，尽管无义多态性对基因产物有巨大影响，但我们对其在人类基因中的分布了解甚少。
方法/主要发现：为了检测影响基因功能的多态性，我们分析了单核苷酸多态性数据库（dbSNP build 125）中所有公开可用的多态性，这些多态性位于36,712个已知和预测的蛋白质编码基因的外显子中，这些基因在所有人类基因和转录本的注释项目（H-InvDB ver3.8）中被定义。我们在这些基因的代表性转录本中总共发现了252,555个单核苷酸多态性（SNP）和8,479个插入和缺失。位于开放阅读框（ORF）中的SNP包括40,484个同义SNP和53,754个非同义SNP，以及1,258个预测为无义SNP或通读SNP。我们估计无义SNP的密度为每位点0.85×10⁻³，低于非同义SNP的密度（每位点2.1×10⁻³）。平均而言，无义SNP位于原始终止密码子上游250个密码子处，替换最常发生在第一个密码子位置。在无义SNP中，581个被预测会导致转录本的无义介导衰变（NMD），从而阻止翻译。我们发现导致NMD的无义SNP在涉及激酶活性和转运的基因中更为常见。其余602个无义SNP预计会产生截短的多肽，平均截短75个氨基酸。此外，还检测到110个终止密码子处的通读SNP。
结论/意义：我们对无义多态性的全面探索表明，无义SNP的密度低于非同义SNP，这表明无义突变比氨基酸变化具有更严重的影响。无义SNP与已知病理变异的对应关系表明，仅一小部分无义SNP的表型效应已被报道，并且导致NMD的无义SNP更有可能参与表型变异。这些无义SNP可能包括尚未报道的病理变异。这些数据可从H-InvDB的转录本视图和VarySysDB（http://h-invitational.jp/varygene/）获得。
