Denault William R P, Romanowska Julia, Helgeland Øyvind, Jacobsson Bo, Gjessing Håkon K, Jugessur Astanand
Department of Genetics and Bioinformatics, Norwegian Institute of Public Health, Oslo, Norway.
Department of Global Public Health and Primary Care, University of Bergen, Bergen, Norway.
BMC Genomics. 2021 May 2;22(1):321. doi: 10.1186/s12864-021-07582-6.
Birth weight (BW) is one of the most widely studied anthropometric traits in humans because of its role in various adult-onset diseases. The number of loci associated with BW has increased dramatically since the advent of whole-genome screening approaches such as genome-wide association studies (GWASes) and meta-analyses of GWASes (GWAMAs). To further contribute to elucidating the genetic architecture of BW, we analyzed a genotyped Norwegian dataset with information on child's BW (N=9,063) using a slightly modified version of a wavelet-based method by Shim and Stephens (2015) called WaveQTL.
WaveQTL uses wavelet regression for regional testing and offers a more flexible functional modeling framework compared to conventional GWAS methods. To further improve WaveQTL, we added a novel feature termed "zooming strategy" to enhance the detection of associations in typically small regions. The modified WaveQTL replicated five out of the 133 loci previously identified by the largest GWAMA of BW to date by Warrington et al. (2019), even though our sample size was 26 times smaller than that study and 18 times smaller than the second largest GWAMA of BW by Horikoshi et al. (2016). In addition, the modified WaveQTL performed better in regions of high LD between SNPs.
This study is the first adaptation of the original WaveQTL method to the analysis of genome-wide genotypic data. Our results highlight the utility of the modified WaveQTL as a complementary tool for identifying loci that might escape detection by conventional genome-wide screening methods due to power issues. An attractive application of the modified WaveQTL would be to select traits from various public GWAS repositories to investigate whether they might benefit from a second analysis.
出生体重（BW）是人类研究最为广泛的人体测量学特征之一，因为它在多种成人期疾病中发挥作用。自全基因组筛查方法（如全基因组关联研究（GWAS）和GWAS的荟萃分析（GWAMA））出现以来，与BW相关的基因座数量急剧增加。为了进一步有助于阐明BW的遗传结构，我们使用Shim和Stephens（2015年）提出的基于小波的方法（称为WaveQTL）的略微修改版本，分析了一个具有儿童BW信息的挪威基因分型数据集（N = 9,063）。
WaveQTL使用小波回归进行区域检测，与传统的GWAS方法相比，提供了更灵活的功能建模框架。为了进一步改进WaveQTL，我们添加了一个名为“缩放策略”的新功能，以增强在通常较小区域中关联的检测。尽管我们的样本量比该研究小26倍，比Horikoshi等人（2016年）的第二大BW的GWAMA小18倍，但修改后的WaveQTL在Warrington等人（2019年）迄今为止最大的BW的GWAMA先前鉴定的133个基因座中的5个上得到了重复。此外，修改后的WaveQTL在SNP之间的高连锁不平衡区域表现更好。
本研究是原始WaveQTL方法首次应用于全基因组基因型数据分析。我们的结果突出了修改后的WaveQTL作为一种补充工具的效用，用于识别由于功效问题可能无法通过传统全基因组筛查方法检测到的基因座。修改后的WaveQTL的一个有吸引力的应用是从各种公共GWAS存储库中选择性状，以研究它们是否可能从二次分析中受益。
