What question did this study set out to answer?

To develop an interpretable machine-learning framework for distinguishing pathogenic from benign missense variants.

March 27, 2026Journal of Genome Biotechnology and GeneticsOpen Access

PathoPredictor: A Machine Learning Framework for Predicting Pathogenic Missense Variants in the Human Genome

View Full Paper

Ask AI

Bookmark

To develop an interpretable machine-learning framework for distinguishing pathogenic from benign missense variants.
Curated clinical variant data and functional annotations were used for model development.
59,302 expert-reviewed missense variants were retained after data processing.
Six machine-learning algorithms were evaluated using cross-validation on the training set.
LightGBM was selected based on its strong performance metrics.
Model interpretability assessed through SHAP for feature ranking and explanation.
LightGBM achieved a mean ROC–AUC of 0.93 ± 0.004 and accuracy of 0.90 ± 0.006.
The model showed a Matthew’s correlation coefficient of 0.80 ± 0.008 across validation folds.
Temporal validation with ClinVar showed consistent ROC–AUC of 0.91 on new submissions.

No takes yet. Share an insight, caveat, or question.

Machine learning framework predicts pathogenic missense variants in genomic data, highlighting its clinical relevance.

Bahmane et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark