Huang Lam Opal, Infante-Rivard Claire, Labbe Aurélie
Section for Metabolic Genetics, Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen, Universitetsparken 1, 1st floor, DIKU, 2100, København Ø, Denmark.
Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, QC, H3A 1A2, Canada.
Hum Genet. 2017 Aug;136(8):951-961. doi: 10.1007/s00439-017-1824-5. Epub 2017 Jun 19.
Transmission ratio distortion (TRD) is a phenomenon where parental transmission of disease allele to the child does not follow the Mendelian inheritance ratio. TRD occurs in a sex-of-parent-specific or non-sex-of-parent-specific manner. An offset computed from the transmission probability of the minor allele in control-trios can be added to the loglinear model to adjust for TRD. Adjusting the model removes the inflation in the genotype relative risk (RR) estimate and Type 1 error introduced by non-sex-of-parent-specific TRD. We now propose to further extend this model to estimate an imprinting parameter. Some evidence suggests that more than 1% of all mammalian genes are imprinted. In the presence of imprinting, for example, the offspring inheriting an over-transmitted disease allele from the parent with a higher expression level in a neighboring gene is over-represented in the sample. TRD mechanisms such as meiotic drive and gametic competition occur in a sex-of-parent-specific manner. Therefore, sex-of-parent-specific TRD (ST) leads to over-representation of maternal or paternal alleles in the affected child. As a result, ST may bias the imprinting effect when present in the sample. We propose a sex-of-parent-specific transmission offset in adjusting the loglinear model to account for ST. This extended model restores the correct RR estimates for child and imprinting effects, adjusts for inflation in Type 1 error, and improves performance on sensitivity and specificity compared to the original model without ST offset. We conclude that to correctly interpret the association signal of an imprinting effect, adjustment for ST is necessary to ensure valid conclusions.
传递比率失真（TRD）是一种现象，即疾病等位基因从父母向子女的传递不遵循孟德尔遗传比率。TRD以亲本性别特异性或非亲本性别特异性的方式发生。可以将根据对照三联体中小等位基因的传递概率计算出的偏移量添加到对数线性模型中，以调整TRD。调整模型可消除由非亲本性别特异性TRD引入的基因型相对风险（RR）估计值的膨胀和I型错误。我们现在提议进一步扩展该模型以估计印记参数。一些证据表明，所有哺乳动物基因中有超过1%是印记基因。例如，在存在印记的情况下，从相邻基因表达水平较高的亲本继承过度传递的疾病等位基因的后代在样本中所占比例过高。减数分裂驱动和配子竞争等TRD机制以亲本性别特异性的方式发生。因此，亲本性别特异性TRD（ST）会导致受影响儿童中母本或父本等位基因的过度代表。结果，当样本中存在ST时，可能会使印记效应产生偏差。我们提议在调整对数线性模型时采用亲本性别特异性传递偏移量来考虑ST。与没有ST偏移量的原始模型相比，这个扩展模型恢复了对儿童和印记效应的正确RR估计，调整了I型错误的膨胀，并提高了敏感性和特异性方面的表现。我们得出结论，为了正确解释印记效应的关联信号，有必要对ST进行调整以确保得出有效的结论。
