Using random forest to detect multiple inherited metabolic diseases simultaneously based on GC-MS urinary...

Area of Science:

• Biochemistry
• Genetics
• Computational Biology

Background:

• Inborn errors of metabolism (IMDs) are genetic disorders causing metabolic disruptions in newborns, potentially leading to severe health issues or death.
• Early and accurate diagnosis of IMDs is crucial for improving newborn survival rates and clinical outcomes.

Purpose of the Study:

• To develop and validate a novel strategy for the simultaneous detection of six common types of IMDs.
• To evaluate the performance of a random forest (RF) algorithm combined with GC-MS for classifying IMDs.

Main Methods:

• Gas chromatography-mass spectrometry (GC-MS) was employed to analyze metabolomics data from urine samples of IMD patients and healthy individuals.
• A random forest (RF) algorithm was developed and trained as a multi-classification tool using the GC-MS data.
• The RF model's performance was compared against artificial neural network (ANN) and support vector machine (SVM) models.

Main Results:

• The RF model demonstrated superior performance in identifying all six types of IMDs and normal samples compared to ANN and SVM models.
• The proposed strategy achieved high specificity, sensitivity, precision, accuracy, and Matthews correlation coefficients.
• GC-MS combined with RF provides a robust method for distinguishing between different IMDs and healthy controls.

Conclusions:

• The integration of GC-MS with the RF algorithm offers a reliable and effective method for the simultaneous identification of multiple IMDs.
• This strategy holds significant potential for improving clinical diagnosis and management of inborn errors of metabolism in newborns.
• The proposed approach provides a valuable tool for timely detection, contributing to better newborn survival and health outcomes.