【佳学基因检测】如何应用大脑影像资料提高神经系统疾病的基因检测正确性？

脑磁共振成像的进展使神经科学有了许多发现。病例组和对照组之间脑部MRI特征的比较突出了精神和行为特征的潜在原因。然而，由于收集MRI数据的成本和招募特定患者群体的困难，大多数研究样本量小，限制了其高效性。此外，反向因果关系使解释变得复杂，许多观察到的大脑差异是疾病的结果，而不是原因。在此，我们提出了一种方法（BrainXcan），该方法利用大规模全基因组关联研究（GWAS）、参考脑MRI数据以及使用遗传仪器进行因果推断的方法学进展，以发现疾病病因的新机制并验证现有机制。BrainXcan测试复杂性状与大脑MRI衍生表型的遗传预测因子的关联，以正确定位相关的区域特异性和跨大脑特征。它还使用孟德尔随机法评估因果流的一致性和方向。由于这种方法只需要遗传数据，BrainXcan允许我们使用现有的公共数据资源，测试关于精神疾病的一系列假设，跨越许多疾病和MRI模式。我们的方法表明，整个大脑的轴突密度降低与精神分裂症的风险相关，这与断连假说一致。我们还发现海马、杏仁核和前扣带回皮质等与精神分裂症风险相关的结构特征，突出了我们的方法带来正交证据的潜力，以告知复杂特征的生物学。
Advances in brain MRI have enabled many discoveries in neuroscience. Comparison of brain MRI features between cases and controls have highlighted potential causes of psychiatric and behavioral traits. However, due to the cost of collecting MRI data and the difficulty in recruiting particular patient groups, most studies have small sample sizes, limiting their reliability. Furthermore, interpretation is complicated by reverse causality, where many observed brain differences are the result of disease rather than the cause. Here we propose a method (BrainXcan) that leverages the power of large-scale genome-wide association studies (GWAS), reference brain MRI data, and methodological advances in causal inference using genetic instruments to discover new mechanisms of disease etiology and validate existing ones. BrainXcan tests complex traits for association with genetic predictors of brain MRI derived phenotypes to pinpoint relevant region-specific and cross-brain features. It also evaluates consistency and direction of the causal flow with Mendelian Randomization. As this approach requires only genetic data, BrainXcan allows us to test a host of hypotheses on mental illness, across many disorders and MRI modalities, using existing public data resources. Our method shows that reduced axonal density across the brain is associated with the risk of schizophrenia, consistent with the disconnectivity hypothesis. We also find structural features hippocampus, amygdala, and anterior cingulate cortex among others associated with schizophrenia risk highlighting the potential of our approach to bring orthogonal lines of evidence to inform the biology of complex traits.

脑磁共振成像的进展使神经科学有了许多发现。然而，在很大程度上，由于样本量小，全脑关联研究（BWAS）的再现性较低（Marek et al.，2020）。这些小样本是由于收集MRI扫描的高成本，以及招募特定精神疾病患者的困难。此外，与疾病状态不会改变种系遗传变异的全基因组关联研究不同，大脑特征可以通过疾病状态和治疗而改变，这可能由于反向因果关系而产生显著关联。

Advances in brain MRI have enabled many discoveries in neuroscience. However, reproducibility of brain-wide associations studies (BWAS) is low due, in large part, to small sample sizes (Marek et al., 2020). These small sample sizes are the result of the high cost of collecting MRI scans, as well as the difficulty in recruiting patients with particular mental illnesses. Also, unlike genome-wide association studies where disease status does not alter germline genetic variation, brain features can be altered by disease status and treatments, which can yield significant associations due to reverse causality.

英国生物银行正在对100000人的脑部MRI进行测量（Littlejohns等人，2020年）。史无前例的数据规模、数据的自动化统一处理、基因数据的可用性和无数的表型数据无疑将促进未来几年的许多发现。对脑MRI图像衍生表型（IDPs）的中期分析发现了许多与之相关的全基因组重要位点，并确定大多数IDPs是可遗传的（Smith等人，2021年）。Zhao等人利用19629名英国生物银行参与者的数据，对101种脑体积表型进行了多基因风险评分，结果表明，他们可以解释四项独立研究中6%以上的表型变异（Zhao等人，2019年）。

The UK Biobank is in the process of measuring brain MRI in 100,000 individuals (Littlejohns et al., 2020). The unprecedented scale of the data, the automated uniform processing of the data, the availability of genetic and a myriad of phenotypic data will undoubtedly catalyze many discoveries in the coming years. The interim analysis of brain MRI image derived phenotypes (IDPs) found many genome-wide significant loci associated and established that most IDPs are heritable (Smith et al., 2021). Zhao et al generated polygenic risk scores of 101 brain volumetric phenotypes using 19,629 UK Biobank participant data and showed they could explain more than 6% of the phenotypic variance in four independent studies (Zhao et al., 2019).

精神病基因组学协会是世界各地研究人员的合作成果，它结合了许多精神疾病的研究，并促成了在每项研究中都不可能实现的发现。他们所有的GWAS总结结果都是公开的，以允许其他研究者测试他们自己的假设并提取新的生物学见解。PGC研究11种精神疾病，包括ADHD、阿尔茨海默病、孤独症、双相情感障碍和精神分裂症。
The Psychiatric Genomics Consortium is a cooperative effort of investigators across the world that combines studies of many mental disorders and has enabled discoveries that would not have been possible within each of the studies. All their GWAS summary results are publicly available to allow other investigators to test their own hypotheses and extract new biological insight. The PGC studies 11 psychiatric disorders including ADHD, Alzheimer’s disease, autism, bipolar disorder, and schizophrenia.(责任编辑：佳学基因)