A considerable proportion of missense variants belong to the “variants of uncertain significance” and therefore pose a challenge for genetic diagnosis (Richards, Aziz, Bale, et al., 2015). Estimation largely relies on calculation‐based tools. However, empirical, experimental, functional, and clinical data are essential. In the case of the NF1 gene which is involved in the genetic disorder neurofibromatosis type 1 (NF1, MIM 162200; Ferner et al., 2007), empirical, genetic and clinical data for pathogenicity for some recurrent missense NF1 variants is accumulating (Jang et al., 2016; Koczkowska et al., 2019; Koczkowska, Callens, Chen, et al., 2020; Pros et al., 2008). In contrast, empirical data for downgrading missense NF1 variants is lacking.

We conceived a strategy to accumulate empirical data for downgrading some missense variants. We searched our dataset for cases with a missense and a co‐existing truncating NF1 variant including nonsense, frameshift and canonical splice variants, deletions of exons, and deletions of the whole NF1 gene. A total of 355 such truncating variants and 72 variants of unknown significance were identified in our cohort.

Among the 72 patients with NF1 variants of unknown significance, 10 had co‐existing truncating NF1 variants, providing empirical supporting data for downgrading these 10 missense variants (Table 1). All these variants were confirmed by repeated Sanger sequencing for the corresponding exons using DNA extracted from different batch blood of the patients.

Theoretically, it is possible that some of these 10 missense NF1 variants are pathogenic and coincidently occurred in the same individual with a truncating variant. However, the probability of such events should be low for the following reasons. First, we are not aware of any such cases in the literatures and databases. Second, bi‐allelic inactivation of the NF1 gene (two pathogenic variants in trans) is lethal as shown in the mouse studies. Third, not a single case with two truncating NF1 variants has been observed in our patient cohort despite the fact that truncating variants are far more frequent than missense variants in the same cohort (355 vs. 72).

A limitation of the present study is the lack of detailed clinical and family data. In addition, we do not know the cis/trans relation of the two variants in each case. We will continue to address these issues in more comprehensive future studies and hope that increasing data will be accumulated from more laboratories and research groups.

In summary, our findings add empirical supportive data for downgrading 10 NF1 variants of unknown significance which may contribute to diagnosis and to improve predicting algorithms.

很大一部分错义变体属于“不确定意义的变体”，因此对遗传诊断提出了挑战（Richards，Aziz，Bale等，2015）。估计在很大程度上依赖于基于计算的工具。但是，经验，实验，功能和临床数据至关重要。对于涉及1型遗传性神经纤维瘤病（NF1，MIM 162200; Ferner等人，2007）的NF1基因，一些复发性错义NF1变异的致病性的经验，遗传和临床数据正在积累（Jang等等人，2016 ; Koczkowska等人，2019 ; Koczkowska，Callens，Chen等人，2020; Pros等，2008）。相反，缺乏用于降级错义NF1变体的经验数据。

我们构想了一种积累经验数据的策略，用于降级某些错义变体。我们在数据集中搜索了具有错义和并存的截断NF1变体的病例，包括无义，移码和规范剪接变体，外显子缺失和整个NF1基因缺失。在我们的队列研究中，总共鉴定出355个这样的截短变体和72个未知意义的变体。

在72例意义不明的NF1变异患者中，有10例同时存在截短的NF1变异，为降级这10个错义变异提供了经验支持数据（表1）。通过使用从患者不同批次血液中提取的DNA对相应外显子重复进行Sanger测序，证实了所有这些变异。

从理论上讲，这10个错义NF1变体中的一些可能是致病性的，并且在具有截断变体的同一个体中同时发生。但是，由于以下原因，此类事件的可能性应较低。首先，我们在文献和数据库中没有发现任何此类情况。其次，如小鼠研究所示，NF1基因的双等位基因失活（两个反式致病变体）具有致命性。第三，在我们的患者队列中未观察到具有两个截短的NF1变体的单个病例，尽管在同一队列中截断变体比错义变体的发生率要高得多（355 vs. 72）。

本研究的局限性是缺乏详细的临床和家庭数据。另外，我们不知道每种情况下两个变体的顺/反关系。我们将在未来更全面的研究中继续解决这些问题，并希望从更多的实验室和研究小组中收集到越来越多的数据。

总而言之，我们的发现为降低10种未知重要性的NF1变体添加了经验支持数据，这可能有助于诊断并改善预测算法。