The GO concentration present in hexose MR mixtures increased with prolonged heating (P < 0.05) ( Fig. 3). In ribose MR mixtures, the GO content increased within 5 min before decreasing with prolonged heating (P < 0.05). This observation can be explained by the fact that the reaction rate of ribose is relatively faster than that of fructose and glucose due to the higher proportion of acyclic form in ribose aqueous solution ( Laroque et al., 2008) and the decreased steric hindrance for pentoses compared with hexoses ( Jing & Kitts, 2004). It is feasible therefore that the GO that was produced was quickly transformed by reacting with other intermediate MRPs derived from the pentose models. The type of amino acid present in the MR did not affect the production of GO in ribose model systems. This observation did not occur when hexose sugars were used in MR models. For example, more GO was generated during the same reaction time (≤ 60 min) in the hexose–lysine than in the hexose–glycine model systems (P < 0.05), because the reaction rate of lysine is relatively faster than that of glycine. Extending the heating time to 90 min however resulted in similar amounts of GO in Glu-Lys and Glu-Gly MR mixtures. A comparison of MR models with different sugars and the same amino acid revealed that more GO was produced in models that contained glucose compared to fructose and ribose. This is probable, since the amount of GO produced was quickly removed in the faster pentose models, in comparison with the relatively slower hexose models. In addition, the production of GO in the aldo-hexose was faster than that of keto-hexose model.