Effects of Maternal Nutrient Restriction and Melatonin Supplementation on Cardiomyocyte Cell Development Parameters Using Machine Learning Techniques

SIMPLE SUMMARY: The major objectives of this study were to examine the effects of maternal feed restriction and melatonin supplementation on fetal cardiomyocyte cell development parameters and predict binucleation and hypertrophy using machine learning techniques using pregnant beef heifers. Interestingly, major findings include evidence that compromised pregnancy in cattle leads to a reduction in the number of cardiomyocytes while melatonin treatment can mitigate some of these disturbances. ABSTRACT: The objective of the current study was to examine the effects of maternal feed restriction and melatonin supplementation on fetal cardiomyocyte cell development parameters and predict binucleation and hypertrophy using machine learning techniques using pregnant beef heifers. Brangus heifers (n = 29) were assigned to one of four treatment groups in a 2 × 2 factorial design at day 160 of gestation: (1) 100% of nutrient requirements (adequately fed; ADQ) with no dietary melatonin (CON); (2) 100% of nutrient requirements (ADQ) with 20 mg/d of dietary melatonin (MEL); (3) 60% of nutrient requirements (nutrient-restricted; RES) with no dietary melatonin (CON); (4) 60% of nutrient requirements (RES) with 20 mg/d of dietary melatonin (MEL). On day 240 of gestation, fetuses were removed, and fetal heart weight and thickness were determined. The large blood vessel perimeter was increased in fetuses from RES compared with ADQ (p = 0.05). The total number of capillaries per tissue area exhibited a nutrition by treatment interaction (p = 0.01) where RES-MEL increased capillary number compared (p = 0.03) with ADQ-MEL. The binucleated cell number per tissue area showed a nutrition by treatment interaction (p = 0.010), where it was decreased in RES-CON vs. ADQ-CON fetuses. Hypertrophy was estimated by dividing ventricle thickness by heart weight. Based on machine learning results, for the binucleation and hypertrophy target variables, the Bagging model with 5 Decision Tree estimators and 3 Decision Tree estimators produced the best results without overfitting. In the prediction of binucleation, left heart ventricular thickness feature had the highest Gin importance weight followed by fetal body weight. In the case of hypertrophy, heart weight was the most important feature. This study provides evidence that restricted maternal nutrition leads to a reduction in the number of cardiomyocytes while melatonin treatment can mitigate some of these disturbances.