ABSTRACT Despite the various existing studies about nonsynonymous single nucleotide polymorphisms (nsSNPs), genome-wide studies based on nsSNPs are rare. NsSNPs alter amino acid sequences, affect protein structure and function, and have deleterious effects. By predicting the deleterious effect of nsSNPs, we determined the total risk score per individual. Additionally, the machine learning technique was utilized to find an optimal nsSNP subset that best explains the complete nsSNP effect. A total of 16,100 nsSNPs were selected as the best representatives among 89,519 regressed nsSNPs. In the gene ontology analysis encompassing the 16,100 nsSNPs, DNA metabolic process, chemokine- and immune-related, and reproduction were the most enriched terms. We expect that our risk score prediction and nsSNP marker selection will contribute to future development of extant genome-wide association studies and breeding science more broadly.

中文翻译：

---

使用全基因组测序数据进行非同义单核苷酸多态性的风险预测和标记选择

摘要 尽管关于非同义单核苷酸多态性 (nsSNPs) 的各种现有研究，基于 nsSNPs 的全基因组研究很少见。NsSNPs 改变氨基酸序列，影响蛋白质结构和功能，并产生有害影响。通过预测 nsSNP 的有害影响，我们确定了每个人的总风险评分。此外，机器学习技术被用来找到一个最佳的 nsSNP 子集，最能解释完整的 nsSNP 效应。在 89,519 个回归的 nsSNPs 中，共有 16,100 个 nsSNPs 被选为最佳代表。在包含 16,100 个 nsSNP 的基因本体分析中，DNA 代谢过程、趋化因子和免疫相关以及繁殖是最丰富的术语。