Routine screening of all newborns for inherited disorders began in the 1960s after the American microbiologist Robert Guthrie, MD, PhD, developed a simple test to identify neonates with phenylketonuria.1 Neonates who tested positive could receive treatment before they became symptomatic. Since then, newborn screening (NBS) has become the standard approach to screening populations for several rare disorders, including inherited metabolic diseases. However, as more and more conditions have been added, priority has been given to improving test performance and reducing false‐positive results. False positives can disrupt parent‐child bonding during the critical first weeks of life and can cause lasting distress for parents.2 Therefore, NBS must ideally have high sensitivity and specificity and a minimal need for manual review. Quality control is an essential component of NBS; for example, the Newborn Screening Quality Assurance Program, administered by the Centers for Disease Control and Prevention, studied samples submitted by 648 laboratories in 85 different countries in 2019.3 NBS has been recommended for 34 health conditions in the United States (although the exact number varies by state and country) and is performed with tandem mass spectrometry (MS/MS).4 NBS by MS/MS has ~99% sensitivity and almost 100% specificity.1, 4 For some disorders, MS/MS has low positive predictive value, and results may be nonspecific. It is imperative that NBS minimizes false positives while identifying true positives.

Rapid advances in next‐generation sequencing technology and computing power have led to the widespread adoption of whole‐exome sequencing (WES) in clinical practice within the last decade. In some countries, WES is now commonly used for the rapid diagnosis of seriously ill children expressing a disease phenotype.5 Although this application presents the opportunity to collect and analyze large amounts of DNA sequence data in the newborn period, there is a significant knowledge gap regarding population‐wide performance characteristics, predictive value, and utility of newborn genomic sequencing.

Now, Adhikari et al6 report on the first comprehensive comparison of WES and established screening technology, MS/MS. In California, between July 2005 and December 2013, the Genetic Disease Screening Program screened dried blood spots from nearly 4.5 million neonates using a multiplex MS/MS platform. The authors obtained a set of 1728 residual, deidentified, archived dried blood spots representing all cases with inborn errors of metabolism. They also obtained selected blood spots that were initially screened as positive from neonates who were later found to be unaffected.

As a primary screen, they analyzed an “exome slice” of 78 genes associated with the 48 inborn errors of metabolism ascertained by NBS in California. Their pipeline correctly identified 571 of 647 Inborn errors of metabolism (IEM)‐affected infants as having a potentially pathogenic IEM genotype, revealing an overall sensitivity of 88%. In the clinically confident subgroup of individuals, their pipeline achieved 93.7% overall sensitivity. Wider WES analysis identified 11 exome‐positive infants for genes unrelated to their IEM. This produced an overall specificity of 98.4%. This would extrapolate to ~8000 false positives among the half a million annual births in California alone. This is far more than the actual 1367 MS/MS false‐positive cases in 2015. Collectively, these data show that, when used alone, sequencing underperforms the classical MS/MS pipeline and misses some affected babies while identifying many healthy neonates as “positive” and targeting them for unnecessary follow‐up testing.

One limitation to the report by Adhikari et al. is the relatively low number of cases studied (1728 vs ~ half million NBS per year in California).6 The sensitivity of their screen varied largely by disorder and performed better for more prevalent IEMs reaching close to 100% sensitivity. Statistical confidence for very rare IEMs would require larger cohort sizes and more data before definitive conclusions can be drawn about the utility of newborn genomic sequencing in NBS.

Abnormal results trigger second‐tier testing critical to distinguishing a false‐positive from a true‐positive result. Second‐tier tests are typically more sensitive and specific than the primary newborn screening assay, but for various reasons, including cost, time, and complexity, they are not suitable to be used as primary screening assays.

Performing newborn genomic sequencing as a second‐tier test when the primary screening results are abnormal could be a cost‐effective alternative to second‐tier biochemical testing. Adhikari et al., therefore, considered WES a reflex follow‐up test for MS/MS‐positive individuals before conducting second‐tier biochemical/clinical studies.6 They found that WES could facilitate rapid and precise clinical resolution for neonates with positive MS/MS on NBS and propose that sequencing can still be useful in cases that look suspicious but were not clearly identified by MS/MS. One has to note, however, that cost and turnaround time, critical concerns in NBS, have not been considered in their study. They found that the turnaround time of WES for critically ill infants ranged from 2 to 3 weeks to less than 24 hours. This finding, and the relatively modest caseload of positive NBS for IEM (~0.3% of births), suggests that WES could become an economical and cost‐effective second‐tier test after a positive MS/MS result. Nonetheless, clinical consideration for individuals with IEM should dictate whether urgent referral after positive MS/MS is required or if it could await sequencing results.

As metabolic specialists, we must emphasize the importance of biochemical testing; elevated metabolites detected by MS/MS are the result of a functional deficit in a pathway regardless of the genes involved, whereas WES, at best, identifies known pathogenic mutations or variants of unknown significance but provides no data on their functional relevance. This gives the classic methodology superiority over genetic techniques, which are currently also slower and more expensive.

In 2020, NBS is mostly focused on inherited metabolic diseases, an emphasis that might change in the near future. If WES becomes the method of choice for other disorders included in NBS, outside the MS/MS panel and performed in every newborn, a combination of the two methods (WES and MS/MS) seems to be a logical option.

In summary, WES alone may not meet standard criteria for NBS yet, but sequencing could be used as a second‐tier test for positive MS/MS results and could reveal a gene variant that provides us with a definite diagnosis, if testing is fast enough and cost‐effective. Several Mendelian conditions, such as neurogenetic disorders with upcoming treatment options, are not amenable to MS/MS and currently go unrecognized until it is too late for optimal intervention. In these cases, NBS by WES could potentially offer an early definitive diagnosis.

1960 年代，美国微生物学家 Robert Guthrie 医学博士开发了一种简单的检测方法来识别患有苯丙酮尿症的新生儿后，开始对所有新生儿进行遗传性疾病的常规筛查。1检测呈阳性的新生儿可以在出现症状之前接受治疗。从那时起，新生儿筛查 (NBS) 已成为筛查多种罕见疾病（包括遗传性代谢疾病）人群的标准方法。然而，随着越来越多的条件被添加，提高测试性能和减少假阳性结果被优先考虑。误报会在生命的最初几周破坏亲子关系，并可能给父母带来持久的痛苦。2因此，理想情况下，NBS 必须具有高灵敏度和特异性，并且最少需要人工审核。质量控制是 NBS 的重要组成部分；例如，由疾病控制和预防中心管理的新生儿筛查质量保证计划研究了 2019 年 85 个不同国家的 648 个实验室提交的样本。3 NBS 已被推荐用于美国的 34 种健康状况（尽管确切的数量因州和国家/地区而异）并使用串联质谱 (MS/MS) 进行。4 MS/MS 的 NBS 具有 ~99% 的灵敏度和几乎 100% 的特异性。1、4对于某些疾病，MS/MS 的阳性预测值较低，结果可能是非特异性的。NBS 必须在识别真阳性的同时最大限度地减少误报。

在过去十年中，新一代测序技术和计算能力的快速进步导致全外显子组测序（WES）在临床实践中得到广泛采用。在一些国家，WES 现在通常用于快速诊断表达疾病表型的重病儿童。5尽管此应用程序提供了在新生儿时期收集和分析大量 DNA 序列数据的机会，但在新生儿基因组测序的全人群表现特征、预测价值和效用方面存在重大的知识差距。

现在，Adhikari 等人6报告了 WES 和已建立的筛选技术 MS/MS 的首次综合比较。在加利福尼亚州，2005 年 7 月至 2013 年 12 月期间，遗传病筛查计划使用多重 MS/MS 平台筛查了近 450 万新生儿的干血斑。作者获得了一组 1728 个残留的、未识别的、存档的干血斑，代表所有先天性代谢错误的病例。他们还从新生儿中获得了最初筛查为阳性的选定血斑，后来发现这些血斑未受影响。

作为初步筛选，他们分析了与加州国家统计局确定的 48 种先天性代谢错误相关的 78 个基因的“外显子组切片”。他们的管道正确地将 647 名受先天性代谢缺陷 (IEM) 影响的婴儿中的 571 名识别为具有潜在致病性 IEM 基因型，显示总体敏感性为 88%。在临床自信的个体亚组中，他们的管道实现了 93.7% 的整体敏感性。更广泛的 WES 分析确定了 11 名外显子组阳性婴儿的基因与其 IEM 无关。这产生了 98.4% 的总体特异性。这将推断仅在加利福尼亚州每年有 50 万新生儿中有大约 8000 个误报。这远远超过 2015 年实际的 1367 个 MS/MS 假阳性案例。 总的来说，这些数据表明，单独使用时，

Adhikari 等人的报告的一个限制。是研究的案例数量相对较少（加利福尼亚州每年 1728 对大约 50 万 NBS）。6其屏幕的灵敏度因无序而有很大差异，并且对于接近 100% 灵敏度的更普遍的 IEM 表现更好。非常罕见的 IEM 的统计置信度需要更大的队列规模和更多的数据才能得出关于新生儿基因组测序在 NBS 中的实用性的明确结论。

异常结果会触发二级测试，这对于区分假阳性和真阳性结果至关重要。二线检测通常比初级新生儿筛查试验更敏感和特异，但由于各种原因，包括成本、时间和复杂性，它们不适合用作初级筛查试验。

当初步筛查结果异常时，将新生儿基因组测序作为二级测试进行可能是二级生化测试的一种经济高效的替代方案。因此，Adhikari 等人在进行二级生化/临床研究之前将 WES 视为 MS/MS 阳性个体的反射性随访测试。6他们发现 WES 可以促进 NBS 上 MS/MS 阳性的新生儿的快速和精确的临床分辨率，并提出测序仍然可以用于看起来可疑但 MS/MS 没有明确识别的病例。然而，必须注意的是，他们的研究并未考虑成本和周转时间，这是 NBS 中的关键问题。他们发现，WES 对危重婴儿的周转时间从 2 到 3 周到不到 24 小时不等。这一发现以及 IEM 的 NBS 阳性病例数相对较少（约占出生人数的 0.3%），表明在 MS/MS 结果呈阳性后，WES 可以成为一种经济且具有成本效益的二级检测。尽管如此，对于 IEM 患者的临床考虑应该决定是否需要在阳性 MS/MS 后紧急转诊，或者是否可以等待测序结果。

作为代谢专家，我们必须强调生化检测的重要性；MS/MS 检测到的代谢物升高是通路功能缺陷的结果，无论涉及的基因如何，而 WES 最多只能识别已知的致病突变或意义未知的变异，但没有提供有关其功能相关性的数据。这使得经典方法优于遗传技术，后者目前也更慢且更昂贵。

2020 年，NBS 主要关注遗传性代谢疾病，这一重点可能会在不久的将来发生变化。如果 WES 成为 NBS 中包含的其他疾病的首选方法，在 MS/MS 面板之外并在每个新生儿中进行，这两种方法（WES 和 MS/MS）的组合似乎是一个合乎逻辑的选择。

总而言之，单独的 WES 可能尚不符合 NBS 的标准，但测序可以用作 MS/MS 结果阳性的二级测试，如果测试速度足够快，可以揭示为我们提供明确诊断的基因变异并且具有成本效益。几种孟德尔疾病，例如具有即将到来的治疗方案的神经遗传性疾病，不适合 MS/MS，并且目前无法识别，直到进行最佳干预为时已晚。在这些情况下，WES 的 NBS 可能会提供早期的明确诊断。