Investigating gene methylation signatures for fetal intolerance prediction
Fig 1
Whole procedures for analyzing the gene expression and methylation profiles of fetal intolerance.
The original dataset is retrieved from Gene Expression Omnibus. Some feature selection methods (Boruta feature filtering and minimum redundancy maximum relevance) follow to analyze the dataset, resulting in a feature list. Incremental feature selection generates several feature subsets from the feature list, on each of which a model is built using one of three classification algorithms. Models are tested by 10-fold cross-validation using synthetic minority oversampling technique to tackle imbalanced problem. Finally, some efficient classifiers, essential genes and decision rules are obtained.