Department of Neurology, Yale School of Medicine, New Haven, CT, USA.
Department of Neurology, Klinikum rechts der Isar, TUM School of Medicine, Technical University of Munich, Ismaninger Str. 22, Munich, Germany.
Commun Biol. 2021 Mar 4;4(1):279. doi: 10.1038/s42003-021-01823-w.
Genetic mapping studies have identified thousands of associations between common variants and hundreds of human traits. Translating these associations into mechanisms is complicated by two factors: they fall into gene regulatory regions; and they are rarely mapped to one causal variant. One way around these limitations is to find groups of traits that share associations, using this genetic link to infer a biological connection. Here, we assess how many trait associations in the same locus are due to the same genetic variant, and thus shared; and if these shared associations are due to causal relationships between traits. We find that only a subset of traits share associations, with many due to causal relationships rather than pleiotropy. We therefore suggest that simply observing overlapping associations at a genetic locus is insufficient to infer causality; direct evidence of shared associations is required to support mechanistic hypotheses in genetic studies of complex traits.
遗传图谱研究已经确定了成千上万种常见变异与数百个人类特征之间的关联。将这些关联转化为机制受到两个因素的影响：它们属于基因调控区域；并且很少映射到一个因果变异上。解决这些限制的一种方法是找到具有共同关联的特征组，利用这种遗传联系来推断生物学联系。在这里，我们评估了同一个基因座上有多少个特征关联是由于相同的遗传变异，因此是共享的；以及这些共享关联是否是由于特征之间的因果关系。我们发现，只有一部分特征具有共享的关联，其中许多是由于因果关系而不是多效性。因此，我们建议，仅仅观察遗传位点上重叠的关联不足以推断因果关系；需要直接证据来支持复杂性状的遗传研究中的机制假说。
