Supplementary Material for: Investigation of recessive effects of coding variants on common clinical phenotypes in exome-sequenced UK Biobank participants

Introduction Previous studies have demonstrated effects of rare coding variants on common, clinically relevant phenotypes although the additive burden of these variants makes only a small contribution to overall trait variance. Although recessive effects of individual homozygous variants have been studied, little work has been done to elucidate the impact of rare coding variants occurring together as compound heterozygotes. Methods In this study attempts were made to identify pairs of variants likely to be occurring as compound heterozygotes using 200,000 exome sequenced subjects from the UK Biobank. Pairs of variants which were seen together in the same subject more often than would be expected by chance were excluded as it was assumed that these might be present in the same haplotype. Attention was restricted to variants with minor allele frequency <= 0.05 and to those predicted to alter amino acid sequence or prevent normal gene expression. For each gene, compound heterozygotes were assigned scores based on the rarity and predicted functional consequences of the constituent variants and the scores were used in a logistic regression analysis to test for association with hypertension, hyperlipidaemia and type 2 diabetes. Results No statistically significant associations were observed and the results conformed to the distribution which would be expected under the null hypothesis. The average number of apparently compound heterozygous subjects for each gene was only 282.2. Conclusion It seems difficult to detect an effect of compound heterozygotes on the risk of these phenotypes. Even if recessive effects from compound heterozygotes do occur they would only affect a small number of people and overall would not make a substantial contribution to phenotypic variance.
This research has been conducted using the UK Biobank Resource.

引言
既往研究已证实，罕见编码变异对常见临床相关表型具有调控效应，但此类变异的加性负荷仅对表型总方差贡献微弱。尽管已有研究针对单个纯合变异的隐性效应展开探讨，但针对以复合杂合子（compound heterozygotes）形式共同存在的罕见编码变异的影响，相关阐释仍较为匮乏。
方法
本研究依托英国生物库（UK Biobank）的20万名接受外显子组测序的受试者数据，旨在筛选出大概率以复合杂合子形式存在的变异对。若某组变异在同一受试者中共同出现的频率高于随机预期水平，则将其排除，因推测此类变异可能位于同一单倍型（haplotype）中。本研究仅纳入次要等位基因频率（minor allele frequency, MAF）≤0.05的变异，以及经预测可改变氨基酸序列或干扰正常基因表达的变异。针对每个基因，研究人员根据组成变异的稀有性与预测的功能效应，为复合杂合子赋值评分，并将评分代入逻辑回归分析，以检验其与高血压（hypertension）、高脂血症（hyperlipidaemia）及2型糖尿病的关联。
结果
本研究未发现具有统计学意义的关联，研究结果符合零假设（null hypothesis）下的预期分布。每个基因对应的疑似复合杂合子受试者平均数量仅为282.2例。
结论
目前难以检测到复合杂合子对上述表型风险的影响。即便复合杂合子确实存在隐性效应，其影响范围也仅局限于少数人群，总体上不会对表型方差产生显著贡献。
本研究依托英国生物库资源开展。