Relative bioavailability of dietary iron sources in growing Japanese quail: Multi-endpoint dose-response and Random...

The present study evaluated the relative bioavailability of Fe supplied as Fe sulfate (FeSO₄), chelated-Fe, nano-Fe oxide, and Fe oxide in growing Japanese quail. A total of 1,200 quails (21 d of age) were allocated to 20 treatments consisting of four Fe sources and five dietary Fe concentrations (30, 60, 90, 120, and 150 mg Fe/kg diet), with five replicate pens of 12 birds each. Growth performance, carcass traits, reproductive organ development, bone characteristics, mineral deposition, and fecal Fe excretion were determined. Relative bioavailability values (RBV) were estimated using linear slope-ratio assays or nonlinear exponential regression models according to the statistical validity of each response criterion. Random Forest analysis was additionally used to identify the biological traits most informative for predicting dietary Fe concentration. Fe source and supplementation level significantly affected body weight gain (P = 0.023 and P < 0.001, respectively) and feed conversion ratio (P = 0.012 and P < 0.001, respectively), with a significant source × level interaction for feed conversion ratio (P = 0.014). Testis weight and tibial Fe concentration were also affected by Fe source, supplementation level, and their interaction (P < 0.05). Chelated-Fe and nano-Fe generally produced greater tibial Fe concentrations than FeSO₄ and Fe oxide, whereas Fe oxide resulted in greater fecal Fe excretion (P < 0.001). Statistical validation supported linear slope-ratio analysis for selected response criteria, including bone density, bone volume, tibial zinc, and fecal Fe, whereas relative bone weight and thigh meat yield showed nonlinear responses. Slope-ratio analysis yielded RBV estimates of 116.6 % for chelated-Fe and 75.3 % for nano-Fe for bone density, while fecal Fe-based estimates were 101.2 % and 84.8 %, respectively, compared with FeSO₄; Fe oxide showed the highest fecal Fe-based estimate (136.3 %). Nonlinear analysis produced numerically favorable RBV estimates for chelated-Fe for relative bone weight (151.2 %) and thigh meat yield (118.0 %), but the associated confidence intervals indicated substantial uncertainty. Random Forest analysis identified FeFecal, relative bone weight, fecal ash, and liver Fe as the most informative predictors of dietary Fe concentration. Overall, Fe source influenced biological responses in an endpoint-dependent manner. Chelated-Fe and nano-Fe showed favorable responses for selected criteria, but the variability and uncertainty of RBV estimates preclude concluding consistent superiority over FeSO₄ across all biological endpoints.