Data_Sheet_1_Multiple-Tissue Integrative Transcriptome-Wide Association Studies Discovered New Genes Associated With Amyotrophic Lateral Sclerosis.docx

Genome-wide association studies (GWAS) have identified multiple causal genes associated with amyotrophic lateral sclerosis (ALS); however, the genetic architecture of ALS remains completely unknown and a large number of causal genes have yet been discovered. To full such gap in part, we implemented an integrative analysis of transcriptome-wide association study (TWAS) for ALS to prioritize causal genes with summary statistics from 80,610 European individuals and employed 13 GTEx brain tissues as reference transcriptome panels. The summary-level TWAS analysis with single brain tissue was first undertaken and then a flexible p-value combination strategy, called summary data-based Cauchy Aggregation TWAS (SCAT), was proposed to pool association signals from single-tissue TWAS analysis while protecting against highly positive correlation among tests. Extensive simulations demonstrated SCAT can produce well-calibrated p-value for the control of type I error and was often much more powerful to identify association signals across various scenarios compared with single-tissue TWAS analysis. Using SCAT, we replicated three ALS-associated genes (i.e., ATXN3, SCFD1, and C9orf72) identified in previous GWASs and discovered additional five genes (i.e., SLC9A8, FAM66D, TRIP11, JUP, and RP11-529H20.6) which were not reported before. Furthermore, we discovered the five associations were largely driven by genes themselves and thus might be new genes which were likely related to the risk of ALS. However, further investigations are warranted to verify these results and untangle the pathophysiological function of the genes in developing ALS.

全基因组关联研究（GWAS）已鉴定出多种与肌萎缩侧索硬化症（ALS）相关的致病基因；然而，ALS的遗传结构尚完全未解，大量致病基因仍有待发现。为填补这一研究空白，本研究对ALS的转录组全关联研究（TWAS）进行了整合分析，以优先考虑由80,610名欧洲个体的汇总统计得到的致病基因，并利用13个GTEx脑组织作为参考转录组面板。首先开展了以单一脑组织为基准的汇总水平TWAS分析，随后提出了名为基于汇总数据的柯西聚合TWAS（SCAT）的灵活p值组合策略，旨在汇总单一组织TWAS分析中的关联信号，同时防止测试间高度正相关性。广泛的模拟实验表明，SCAT能够生成良好的校准p值，以控制I型错误，并且在识别各种场景下的关联信号方面，通常比单一组织TWAS分析具有更高的功效。利用SCAT，我们验证了先前在GWAS中发现的三个与ALS相关的基因（即ATXN3、SCFD1和C9orf72），并发现了五个先前未报道的基因（即SLC9A8、FAM66D、TRIP11、JUP和RP11-529H20.6）。此外，我们发现这五种关联主要是由基因自身驱动的，因此可能是与ALS风险相关的新的基因。然而，进一步的调查研究对于验证这些结果并阐明基因在ALS发病过程中的病理生理功能是必要的。