Department of Biostatistics, Epidemiology and Informatics, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA 19104, USA.
Pac Symp Biocomput. 2022;27:109-120.
Brain imaging genetics, an emerging and rapidly growing research field, studies the relationship between genetic variations and brain imaging quantitative traits (QTs) to gain new insights into the phenotypic characteristics and genetic mechanisms of the brain. Heritability is an important measurement to quantify the proportion of the observed variance in an imaging QT that is explained by genetic factors, and can often be used to prioritize brain QTs for subsequent imaging genetic association studies. Most existing studies define regional imaging QTs using predefined brain parcellation schemes such as the automated anatomical labeling (AAL) atlas. However, the power to dissect genetic underpinnings under QTs defined in such an unsupervised fashion could be negatively affected by heterogeneity within the regions in the partition. To bridge this gap, we propose a novel method to define highly heritable brain regions. Based on voxelwise heritability estimates, we extract brain regions containing spatially connected voxels with high heritability. We perform an empirical study on the amyloid imaging and whole genome sequencing data from a landmark Alzheimer's disease biobank; and demonstrate the regions defined by our method have much higher estimated heritabilities than the regions defined by the AAL atlas. Our proposed method refines the imaging endophenotype constructions in light of their genetic dissection, and yields more powerful imaging QTs for subsequent detection of genetic risk factors along with better interpretability.
脑影像遗传学是一个新兴且快速发展的研究领域，它研究遗传变异与脑影像定量性状（QTs）之间的关系，以深入了解大脑的表型特征和遗传机制。遗传力是衡量遗传因素解释观察到的影像 QT 变异比例的重要指标，通常可用于优先选择后续影像遗传关联研究的脑 QT。大多数现有研究使用预定义的脑分割方案（如自动解剖标记 (AAL) 图谱）来定义区域影像 QT。然而，以这种非监督方式定义 QT 时，对遗传基础进行剖析的能力可能会因分区内的异质性而受到负面影响。为了弥合这一差距，我们提出了一种定义高度遗传脑区的新方法。基于体素遗传力估计，我们提取包含高遗传力的空间连接体素的脑区。我们对来自一个有里程碑意义的阿尔茨海默病生物库的淀粉样蛋白影像和全基因组测序数据进行了实证研究；并证明了我们方法定义的区域比 AAL 图谱定义的区域具有更高的遗传力估计值。我们提出的方法根据遗传剖析来细化影像内表型构建，为后续检测遗传风险因素提供更强大的影像 QT，同时具有更好的可解释性。
