生物测量的高斯表型
The Gaussian phenotype of biological measurements
浏览论文内容
中文总结 AI 辅助
研究探讨生物测量的高斯性，用QQ-RMSE量化，通过NHANES数据构建参考分布比较高斯表型，发现不同生物测量有不同高斯特征，生物标准化可改善，提出高斯表型概念并建立参考图谱。
中文摘要 AI 辅助
生物测量通常被认为近似高斯分布，正态性常被作为统计分析前提进行评估。然而，高斯性程度本身是否包含生物信息尚待探索。本文用正态分位数-分位数图的均方根误差（QQ-RMSE）量化生物测量的高斯性。通过1999 - 2023年美国国家健康与营养检查调查（NHANES）的10249次生物测量构建参考分布，比较生物测量的高斯表型。结果显示不同生物测量有特征性高斯表型，生物标准化可改善高斯性。这些发现表明高斯性是可测量的生物属性，本文提出高斯表型概念并建立了首个高斯性参考图谱用于解释生物测量。
英文摘要
Biological measurements are commonly assumed to approximate Gaussian distributions, and normality is routinely assessed as a prerequisite for statistical analysis. However, whether the degree of Gaussianity itself contains biological information remains largely unexplored. Here, we quantified the Gaussianity of biological measurements using the root mean square error of normal quantile-quantile plots (QQ-RMSE). A reference distribution was constructed from 10,249 biological measurements from the National Health and Nutrition Examination Survey (NHANES) 1999-2023, enabling direct comparison of the Gaussian phenotype, defined as the degree to which a biological measurement approximates a Gaussian distribution. Biological measurements exhibited characteristic Gaussian phenotypes. Structural and capacity-related traits, including body measurements, grip strength, spirometry, and red blood cell count, consistently showed low QQ-RMSE values. Homeostatically regulated variables, such as total cholesterol, also exhibited high Gaussianity. In contrast, biomarkers associated with physiological responses or pathology, including triglycerides, C-reactive protein, liver enzymes, serum creatinine, and urinary albumin, showed progressively larger deviations from Gaussianity. Biological normalization further improved Gaussianity: the albumin-to-creatinine ratio consistently exhibited lower QQ-RMSE values than urinary albumin alone across all NHANES survey cycles. These findings indicate that Gaussianity is not merely a statistical assumption but a measurable biological property. We propose the concept of the Gaussian phenotype, in which the degree of Gaussianity reflects biological mechanisms governing variability. This study establishes the first reference atlas of Gaussianity for interpreting biological measurements.
