Correction to: Predicting metabolic responses in genetic disorders via structural representation in machine learning

In this article, the header “Binary Morgan with added properties” in Table 2 has been incorrectly processed as “Binary Morgan” during the production process. The correct Table 2 should be as given below, and the original article has been corrected. The table presents the performance of ML models trained using state-of-the-art molecular structure encoding techniques Binary Morgan Metric A BA ROC MCC F1 R P Bernoulli Naive Bayes 0.794 0.693 0.693 0.333 0.659 0.693 0.644 Gaussian Naive Bayes 0.778 0.690 0.690 0.321 0.647 0.690 0.637 Logistic Regression 0.761 0.697 0.697 0.318 0.641 0.697 0.628 Decision Tree 0.754 0.653 0.653 0.258 0.616 0.653 0.609 Random Forest 0.824 0.682 0.682 0.350 0.673 0.682 0.668 XGBoost 0.794 0.677 0.677 0.311 0.650 0.677 0.637 Support Vector 0.788 0.707 0.707 0.348 0.662 0.707 0.646 MultiLayer Perceptron 0.813 0.591 0.591 0.207 0.599 0.591 0.618 Count Morgan Metric A BA ROC MCC F1 R P Bernoulli Naive Bayes 0.794 0.693 0.693 0.333 0.659 0.693 0.644 Gaussian Naive Bayes 0.794 0.639 0.639 0.275 0.624 0.639 0.643 Logistic Regression 0.779 0.699 0.699 0.330 0.653 0.699 0.637 Decision Tree 0.755 0.640 0.640 0.238 0.608 0.640 0.601 Random Forest 0.825 0.686 0.686 0.354 0.675 0.686 0.670 XGBoost 0.804 0.689 0.689 0.337 0.663 0.689 0.650 Support Vector 0.815 0.681 0.681 0.336 0.666 0.681 0.656 MultiLayer Perceptron 0.809 0.600 0.600 0.216 0.606 0.600 0.617 Binary Morgan with added properties Metric A BA ROC MCC F1 R P Bernoulli Naive Bayes 0.794 0.693 0.693 0.333 0.659 0.693 0.644 Gaussian Naive Bayes 0.767 0.699 0.699 0.326 0.645 0.699 0.634 Logistic Regression 0.759 0.702 0.702 0.324 0.642 0.702 0.630 Decision Tree 0.775 0.723 0.723 0.362 0.661 0.723 0.647 Random Forest 0.700 0.700 0.380 0.688 0.700 XGBoost 0.809 0.672 Support Vector 0.786 0.709 0.709 0.348 0.662 0.709 0.645 MultiLayer Perceptron 0.815 0.595 0.595 0.215 0.603 0.595 0.623 Graph Metric A BA ROC MCC F1 R P GCN (pretrained) 0.722 0.680 0.680 0.283 0.610 0.680 0.612 GCN 0.718 0.675 0.675 0.272 0.605 0.675 0.606 GAT 0.691 0.688 0.688 0.282 0.595 0.688 0.606 GIN 0.721 0.669 0.669 0.268 0.605 0.669 0.607 SAGE 0.742 0.689 0.689 0.301 0.625 0.689 0.621 Reference metrics have been calculated for each model. A bold value in a column indicates that the metric of the corresponding model is better than all the other models for a given structural encoding, with statistical significance determined via a paired t-test at a significance level of 0,05. An underlined value indicates that the metric is significantly better across all models and structural encodings