Machine learning classification models for fetal skeletal development performance prediction using maternal bone...

Area of Science:

• Animal models
• Biochemistry
• Genetics

Background:

• Maternal undernutrition is a major factor affecting fetal development and pregnancy outcomes in developing nations.
• Maternal nutrition restriction (MNR) during gestation affects offspring's growth, bone development, and mesenchymal stem cell metabolism.
• Efficient methods to assess fetal bone development using maternal bone metabolic markers are needed.

Purpose of the Study:

• To establish an efficient method for elucidating fetal bone development using maternal serum bone metabolic markers under malnutrition conditions.
• To utilize machine learning to predict fetal bone parameters based on maternal metabolic data and experimental conditions.
• To identify key maternal bone metabolic markers influencing fetal bone growth.

Main Methods:

• Goats were used to study fetal bone development under maternal malnutrition in mid- and late-gestation.
• Seventy-two datasets were created by combining experimental conditions and metabolic data.
• Seven machine learning methods were employed to predict six fetal bone parameters (femur/humerus weight, length, diameter).

Main Results:

• Maternal nutrition restriction significantly influenced fetal bone development and maternal bone metabolic protein levels (CTx, NTx, BALP).
• Machine learning models, particularly Support Vector Machines, achieved high accuracy (up to 1.0) in predicting fetal bone parameters.
• Feature importance analysis highlighted the role of experimental conditions and specific metabolites like parathyroid hormone (PTH) and bone alkaline phosphatase (BALP).

Conclusions:

• Machine learning provides an efficient approach to confirm the role of maternal metabolic markers (PTH, BALP) and nutritional conditions in fetal bone growth.
• The study demonstrates the potential of using maternal serum markers and machine learning for non-invasive assessment of fetal bone development.
• The findings offer insights into mitigating adverse effects of maternal undernutrition on fetal skeletal health.