PDF(Pubmed)
Abstract：
Pharmacogenomics (PGx) investigates the influence of genetics on drug responses, enabling tailored treatments for personalized healthcare. This study assessed the accuracy of genotyping six genes using whole genome sequencing with four different computational tools and various sequencing depths. The effects of using different reference genomes (GRCh38 and GRCh37) and sequence aligners (BWA-MEM and Bowtie2) were also explored. The results showed generally minor variations in tool performance across most genes; however, more notable discrepancies were observed in the analysis of the complex CYP2D6 gene. Cyrius, a CYP2D6-specific tool, demonstrated the most robust performance, achieving the highest concordance rates for CYP2D6 in all instances, comparable to the consensus approach in most cases. There were rather small differences between the samples with 20× coverage depth and those with higher depth, but the decreased performance was more evident at lower depths, particularly at 5×. Additionally, variations in CYP2D6 results were observed when samples were aligned to different reference genomes using the same method, or to the same genome using different aligners, which led to reporting incorrect rare star alleles in several cases. These findings inform the selection of optimal PGx tools and methodologies as well as suggest that employing a consensus approach with two or more tools might be preferable for certain genes and tool combinations, especially at lower sequencing depths, to ensure accurate results. Additionally, we show how the upstream alignment can affect the performance of tools, an important factor to take into account.
摘要：
药物基因组学（PGx）研究遗传学对药物反应的影响，为个性化医疗保健提供量身定制的治疗方法。这项研究评估了使用四种不同的计算工具和各种测序深度的全基因组测序对六种基因进行基因分型的准确性。还探索了使用不同参考基因组(GRCh38和GRCh37)和序列比对(BWA-MEM和Bowtie2)的效果。结果表明，大多数基因的工具性能通常存在较小的差异；然而，在复杂CYP2D6基因的分析中观察到更显著的差异.Cyrius,CYP2D6专用工具，展示了最强大的性能，在所有情况下实现CYP2D6的最高一致率，在大多数情况下，与共识方法相当。具有20倍覆盖深度的样本与具有较高深度的样本之间存在相当小的差异，但是在较低的深度表现下降更明显，特别是在5×此外,当使用相同的方法将样品与不同的参考基因组比对时，观察到CYP2D6结果的变化，或者使用不同的对齐器对相同的基因组，这导致在一些情况下报告不正确的罕见恒星等位基因。这些发现为选择最佳的PGx工具和方法提供了信息，并表明采用两种或多种工具的共识方法对于某些基因和工具组合可能更可取。尤其是在较低的测序深度，确保结果准确。此外,我们展示了上游对齐如何影响工具的性能，一个需要考虑的重要因素。
