臺灣族群中罕見外顯子變異 對於雙極性情感疾患的遺傳傾向

Abstract

雙極性情感疾患（躁鬱症）的特徵是躁症和鬱症交互發作，大約影響全球1-2%的人口。早期雙胞胎研究所估計遺傳力約為70%，然而，透過全基因組關聯研究所發現的常見變異所貢獻的遺傳力僅為25%。這種差距被稱為「遺傳性缺失」，部分原因可能是 GWAS 未完全捕捉到罕見變異的影響。目前罕見變異的研究結果異質性很高，且關注東亞人群的研究相對稀少。因此，我們的目標是使用全外顯子定序資料去衡量罕見變異對於躁鬱症在臺灣族群的影響。 我們使用透過混合基因外險子組方法產生的WES 數據，分析483名躁鬱症患者和535名對照組。經過一系列的品質管制，我們保留了1,017個個體，並提取出平均深度為40x的外顯子區域，再對其變異使用Ensembl’s Variants Effect Predictor 進行註釋，將其根據預測的功能分為同義變異、錯義變異和蛋白質截斷變異。接著，對於不同次要等位基因個數進行三種負擔測試，以檢查 躁鬱症和罕見變異之間的關聯，包括透過 Firth羅吉斯迴歸分析的全外顯子組和基因集負擔測試，以及使用費雪精確檢定分析的基因負擔測試。最後，採用整合罕見變異攜帶者狀態和躁鬱症多基因風險評分的聯合模型，以探討常見和罕見變異對於躁鬱症遺傳風險的綜合影響。 負擔測試顯示罕見變異在外顯子組範圍內有小而顯著的富集，特別是在較低的次要等位基因個數閾值和對於有嚴重功能的變異類別，例如破壞性錯義變異，而在同義變異中並沒有觀察到效益。當縮小限制在高機率不耐受功能喪失的受限基因時，負擔訊號變得更強。然而，我們沒有觀察到蛋白質截斷變異的效應，這可能因為蛋白質截斷變異的數量過少。儘管沒有發現任何顯著的單一基因，但在與躁鬱症、晝夜節律和電壓門控制離子通道活性相關的基因集中觀察到罕見變異的富集; 也觀察到與精神分裂症和自閉症有關的基因存在關聯性，凸顯了特定的生物學途徑。最後，我們的聯合模型方法顯示，與中等多基因風險評分的蛋白質截斷變異非攜帶者相比，高多基因風險評分的蛋白質截斷變異非攜帶者較有可能罹患躁鬱症，且可能性又略高於高多基因風險評分的蛋白質截斷變異攜帶者。此外，我們也觀察到常見和罕見變異之間可能存在交互作用，代表他們對於躁鬱症有綜合的貢獻。 儘管受到小樣本的限制，我們使用臺灣世代研究中的外顯子數據檢測到罕見變異對於躁鬱症具有中等的影響。此外，我們觀察到躁鬱症和其他精神疾病之間具有重疊的罕見變異負擔，以及常見變異和罕見變異對於躁鬱症的聯合效應。這些發現提供了對東亞人躁鬱症複雜性的見解，並強調了全面性的遺傳分析對於躁鬱症病因的重要性。未來的方向可整合來自不同來源的數據，以擴大樣本數量，並將分析其他更多與躁鬱症相關的特徵。

Bipolar disorder (BD) is characterized by manic and depressive episodes and affects approximately 1-2% of the global population. The heritability from twin studies is about 70%; however, the estimated SNP-based heritability of common variants found by genome-wide association studies (GWAS) is only 25%. This gap is called “missing heritability”, which may be partially explained by the effects of rare deleterious variants not fully captured by GWAS. Research on rare variants has shown heterogenous findings, and few have focused on the East Asian populations. Here, we aim to evaluate the impact of rare deleterious variants on BD in the Taiwanese population using whole exome sequencing (WES) data. We analyzed 483 BD patients and 535 controls in our WES data, generated via the Blended Genome Exome (BGE) method. After a series of quality control, we kept 1,017 individuals and focused on exome regions where the average read depth was 40x. We used Ensembl’s Variant Effect Predictor (VEP) to annotate variants and grouped them into synonymous, missense, and protein-truncating variants (PTVs) based on the predicted functional consequences. Then, we performed three levels of burden tests for qualifying variants at different minor allele count (MAC) thresholds to examine the association between BD and rare deleterious variants, including exome-wide and gene-set burden tests via Firth’s logistic regression, and gene-based burden tests using Fisher’s exact test. Finally, we employed a joint modeling approach integrating rare variants carrier status and BD polygenic risk scores (PRS) to elucidate the combined effect of common and rare genetic variants on BD risk. Burden tests revealed a small yet significant exome-wide enrichment of rare deleterious variants, particularly for more severe functional classes such as damaging missense variants and at lower MAC thresholds, with no inflation observed in synonymous variants. The burden signals became stronger when restricting to constrained genes with high probability of being Loss-of-function Intolerant (pLI scores> 0.95). However, we did not observe an excess of PTVs in BD, likely due to extremely low counts. While no gene emerged as exome-wide significant, an enrichment of rare deleterious variants was observed in genes associated with BD, circadian rhythm and voltage-gated ion channel activity; a suggestive association was observed for gene implicated in schizophrenia and autism spectrum disorder, highlighting specific biological pathways. Finally, our joint model approach showed that comparing to PTVs non-carriers with intermediate PRS, those with high PRS were more likely to develop BD, slightly larger than that among carriers with high PRS. This model also revealed a marginal interaction between common and rare variants, suggesting their combined contribution to BD risk. Despite the constraint of a small sample size, we detected an overall moderate effect of rare variants on BD using exome data in a Taiwanese cohort. Additionally, we observed overlap of rare-variant burden between BD and other psychiatric traits, and a joint effect of common and rare variants underlying BD susceptibility. These findings provide insights into the genetic complexity of BD among East Asians and highlighting the importance of comprehensive genetic analyses to unravel its etiology. Future directions could include integrating data from various sources to enlarge the sample size and expanding the analysis to BD spectrum phenotypes.