Department of Computer Science, Trinity College of Arts and Sciences (R.L.D., F.Z.), Duke University.
Medical Scientist Training Program (R.L.D.), Duke University School of Medicine, Durham, NC.
Circ Arrhythm Electrophysiol. 2022 Apr;15(4):e010326. doi: 10.1161/CIRCEP.121.010326. Epub 2022 Mar 31.
Cardiac channelopathies such as catecholaminergic polymorphic tachycardia and long QT syndrome predispose patients to fatal arrhythmias and sudden cardiac death. As genetic testing has become common in clinical practice, variants of uncertain significance (VUS) in genes associated with catecholaminergic polymorphic ventricular tachycardia and long QT syndrome are frequently found. The objective of this study was to predict pathogenicity of catecholaminergic polymorphic ventricular tachycardia-associated VUS and long QT syndrome-associated VUS in , , and by developing gene-specific machine learning models and assessing them using cross-validation, cellular electrophysiological data, and clinical correlation.
The GENe-specific EnSemble grId Search framework was developed to identify high-performing machine learning models for , , , and using variant- and protein-specific inputs. Final models were applied to datasets of VUS identified from ClinVar and exome sequencing. Whole cell patch clamp and clinical correlation of selected VUS was performed.
The GENe-specific EnSemble grId Search models outperformed alternative methods, with area under the receiver operating characteristics up to 0.87, average precisions up to 0.83, and calibration slopes as close to 1.0 (perfect) as 1.04. Blinded voltage-clamp analysis of HEK293T cells expressing 2 predicted pathogenic variants in each revealed an ≈80% reduction of peak Kv7.1 current compared with WT. Normal Kv7.1 function was observed in KCNQ1-V241I HEK cells as predicted. Though predicted benign, loss of Kv7.1 function was observed for KCNQ1-V106D HEK cells. Clinical correlation of 9/10 variants supported model predictions.
Gene-specific machine learning models may have a role in post-genetic testing diagnostic analyses by providing high performance prediction of variant pathogenicity.
儿茶酚胺多形性室性心动过速和长 QT 综合征等心脏通道病使患者容易发生致命性心律失常和心脏性猝死。随着基因检测在临床实践中的普及，与儿茶酚胺多形性室性心动过速和长 QT 综合征相关的基因中的不确定意义变异（VUS）经常被发现。本研究的目的是通过开发基因特异性机器学习模型，并通过交叉验证、细胞电生理数据和临床相关性来评估这些模型，从而预测儿茶酚胺多形性室性心动过速相关 VUS 和长 QT 综合征相关 VUS 在 、 、 和 中的致病性。
使用变体和蛋白质特异性输入，开发了 GENe-specific EnSemble grId Search 框架，以识别用于 、 、 、 和 的高性能机器学习模型。最终模型应用于 ClinVar 和外显子组测序中确定的 VUS 数据集。对选定的 VUS 进行全细胞膜片钳和临床相关性分析。
GENe-specific EnSemble grId Search 模型的表现优于其他方法，其接收者操作特征曲线下面积高达 0.87，平均精度高达 0.83，校准斜率接近 1.0（完美），为 1.04。对表达 2 种每种预测致病性变异的 HEK293T 细胞进行盲电压钳分析，发现与 WT 相比，Kv7.1 电流峰值降低约 80%。预测 KCNQ1-V241I HEK 细胞中 Kv7.1 功能正常。虽然预测为良性，但观察到 KCNQ1-V106D HEK 细胞中 Kv7.1 功能丧失。9/10 个变体的临床相关性支持模型预测。
基因特异性机器学习模型通过提供对变异致病性的高性能预测，可能在基因检测后诊断分析中发挥作用。
