KCNQ1

KCNQ1 predicts the functional impact of KCNQ1 missense variants, including variants of unknown significance, to support interpretation of genetic testing for long-QT syndrome.

Key Features:

• Data-Driven Approach: Trained on a curated dataset of 107 experimentally characterized KCNQ1 variants.
• Sequence Conservation Analysis: Quantitatively analyzes sequence conservation across KCNQ1 subdomains and identifies conserved regions enriched for dysfunctional variants critical for channel function.
• Neural Network Model: Employs an artificial neural network specifically trained to predict the functional impact of KCNQ1 variants of unknown significance (VUS) and tailored to KCNQ1 characteristics.
• Performance Metrics: Reports a Matthew's correlation coefficient of 0.581 and an area under the receiver operating characteristic curve (AUC) of 0.884, outperforming eight previously tested methods.

Scientific Applications:

• Clinical Genetic Testing: Provides variant impact predictions to inform interpretation of genetic tests for long-QT syndrome.
• Research on KCNQ1 Variants: Supports identification of functionally critical KCNQ1 subdomains and investigation of pathogenic mechanisms of KCNQ1 mutations.

Methodology:

Training of an artificial neural network on a dataset of experimentally validated KCNQ1 variants with incorporation of sequence conservation data to identify patterns associated with dysfunctional variants.