Comparison of analytical and predictive methods for water, protein, fat, sugar, and gross energy in marine mammal milk

摘要: Abstract Mammalian milks may differ greatly in composition from cow milk, and these differences affect the performance of analytical methods. High-fat, high-protein with a preponderance oligosaccharides, such as those produced by many marine mammals, present particular challenge. We compared several methods against reference procedures using Weddell seal ( Leptonychotes weddellii ) milk highly varied (by methods: 27–63% water, 24–62% fat, 8–12% crude protein, 0.5–1.8% sugar). A microdrying step preparatory to carbon-hydrogen-nitrogen (CHN) gas analysis slightly underestimated water content had higher repeatability relative standard deviation (RSD r than did oven drying at 100°C. Compared macro-Kjeldahl protein procedure, CHN (or Dumas) combustion method somewhat RSD (1.56 vs. 0.60%) but correlation between was high (0.992), means were not different (CHN: 17.2±0.46% dry matter basis; Kjeldahl 17.3±0.49% basis), there no significant proportional or constant errors, predictive high. carbon stoichiometric procedure based on failed adequately predict fat (reference: Rose-Gottlieb method) total sugar phenol-sulfuric acid method). Gross energy content, calculated energetic factors results for sugar, accurately predicted gross measured bomb calorimetry. conclude that (Dumas) calculation are acceptable approaches mammal require separate appropriate analytic cannot be estimated stoichiometry. Some other alternative methods—low-temperature determination; Bradford, Lowry, biuret protein; Folch Bligh Dyer fat; enzymatic reducing sugar—appear likely produce substantial error milks. It is important properly validated before being used, especially mammalian analyte characteristics concentrations.