We read with interest the publication by Sándor and Sahin-Tóth, which classifies CPA1 missense variants based on functional data.1 Previous analyses of variants strongly associated with chronic pancreatitis (p.N256K, p.S282P and p.K374E) have demonstrated that pathogenic CPA1 variants result in proenzyme misfolding, leading to subsequent endoplasmic reticulum (ER) stress.2–6 In examining 50 CPA1 missense variants, Sándor and Sahin-Tóth revealed an inverse relationship between BiP mRNA expression, a marker of ER stress, in transfected HEK293T cells, and secretion levels. From these insights, they proposed a classification of CPA1 variants according to secretion levels: pathogenic (<10% of normal secretion), uncertain (10–20%) and benign (>20%). In silico prediction tools are pivotal for variant classification. AlphaMissense, a recent innovation, categorises missense variants across the human proteome into pathogenic, benign or ambiguous groups.7 Its predictions, primarily relying on structural context through an AlphaFold-based framework, substantially decrease dependence on human-curated data. AlphaMissense assigns pathogenicity scores ranging from 0 to 1, where higher scores suggest a greater likelihood of pathogenicity. In this study, we explored the capabilities of …

中文翻译：

---

AlphaMissense 在 CPA1 错义变异分类中的优势和局限性


我们饶有兴趣地阅读了 Sándor 和 Sahin-Tóth 的出版物，该出版物根据功能数据对 CPA1 错义变异进行了分类。先前对与慢性胰腺炎密切相关的变异（p.N256K、p.S282P 和 p.K374E）的分析表明，致病性 CPA1 变异导致酶原错误折叠，从而导致随后的内质网 （ER） 应激。在检查 50 个 CPA1 错义变异时， Sándor 和 Sahin-Tóth 揭示了转染 HEK293T 细胞中 BiP mRNA 表达（ER 应激的标志物）与分泌水平呈负相关关系。根据这些见解，他们提出了根据分泌水平对 CPA1 变体进行分类：致病性 （<10% 正常分泌）、不确定 （10-20%） 和良性 （>20%）。计算机预测工具对于变体分类至关重要。AlphaMissense 是最近的一项创新，它将人类蛋白质组中的错义变异分为致病性、良性或模棱两可的组7。其预测主要依赖于通过基于 AlphaFold 的框架的结构背景，大大减少了对人工整理数据的依赖。AlphaMissense 分配的致病性分数范围为 0 到 1，其中分数越高表示致病性的可能性越大。在这项研究中，我们探讨了 ...