FIGURE 1.

Illustration of machine learning pipeline in NBS. The classification model is the essential part of the ML pipeline in NBS including the interpretable and noninterpretable classification methods and their performance optimization. Data preprocessing is an optional module applied before the classification model. It can include data sampling, feature selection and feature construction methods. Pattern recognition is applied after the classification method evaluating feature importance for biomarker discovery. ML, machine learning; NBS, newborn screening