Automated Clinical Exome Reanalysis Reveals Novel Diagnoses
Automated Clinical Exome Reanalysis Reveals Novel Diagnoses
复制标题
DOI：
10.1016/j.jmoldx.2018.07.008
复制
发表时间：
2019-01-01
影响因子：
4.1
通讯作者：
Santani, Avni B.
中科院分区：
文献类型：
--
作者：
Baker, Samuel W.;Murrell, Jill R.;Santani, Avni B.
临床外显子组测序(CES)对大多数临床适应症的诊断效率为20%至30%。正在进行的新基因疾病和变异疾病关联的发现有望增加CES的诊断效率。对以前未诊断的CES样本进行系统的重新分析对临床实验室来说是一个重大的挑战。在这里，我们展示了一种新的自动再分析方法的结果，该方法应用于2014年6月至2016年9月期间初步分析的300个CES样本。我们的再分析方法的应用减少了93%的再分析变异分析负担，并正确地捕获了60个样本中70个先前识别的诊断变异中的70个。值得注意的是，使用2017年7月1日提供的信息对240个最初非诊断性样本进行了重新分析，发现了38个新诊断，诊断率提高了15.8%。对240个非诊断性样本的建模每月迭代再分析显示，每月样本的诊断率为0.57%。对非诊断性样本的每月迭代再分析所需的工作量进行建模显示，仅先证者样本的变异分析负担约为每月5个变异，三个样本的变异分析负担约为每月0.5个变异。大约45%的样本需要在每个月的间隔期间进行评估，61.3%的样本在连续三次重新分析中进行了重新评估。总之，自动化再分析方法可以利用最新的文献有效地重新评估非诊断样本，并能为临床实验室提供重要的价值。
Clinical exome sequencing (CES) has a reported diagnostic yield of 20% to 30% for most clinical indications. The ongoing discovery of novel gene disease and variant disease associations are expected to increase the diagnostic yield of CES. Performing systematic reanalysis of previously non diagnostic CES samples represents a significant challenge for clinical laboratories. Here, we present the results of a novel automated reanalysis methodology applied to 300 CES samples initially analyzed between June 2014 and September 2016. Application of our reanalysis methodology reduced reanalysis variant analysis burden by >93% and correctly captured 70 of 70 previously identified diagnostic variants among 60 samples with previously identified diagnoses. Notably, reanalysis of 240 initially nondiagnostic samples using information available on July 1, 2017, revealed 38 novel diagnoses, representing a 15.8% increase in diagnostic yield. Modeling monthly iterative reanalysis of 240 non diagnostic samples revealed a diagnostic rate of 0.57% of samples per month. Modeling the workload required for monthly iterative reanalysis of nondiagnostic samples revealed a variant analysis burden of approximately 5 variants/month for proband-only and approximately 0.5 variants/month for trio samples. Approximately 45% of samples required evaluation during each monthly interval, and 61.3% of samples were reevaluated across three consecutive reanalyses. In sum, automated reanalysis methods can facilitate efficient reevaluation of nondiagnostic samples using up-to-date literature and can provide significant value to clinical laboratories.