Assessment of objective texture measurements for characterising and predicting the sensory quality of squash (Cucurbita maxima)

Ten squash (Cucurbita maxima) cultivars were assessed using a trained sensory panel and a range of objective measures. Significant differences between cultivars were found for most sensory, Instron texture profile analysis (TPA), and biochemical properties. Partial least squares (PLS) regression models using single objective TPA variables explained up to 75.7% of variance in squash textural attributes, whereas including multiple objective TPA variables and/or dry matter (DM) increased the variation explained to varying degrees. Sugars and DM explained up to 86.1% and 65.7% of variance in sweetness and flavour intensity ratings respectively. Principal components analysis (PCA) showed that the first two principle components explained 97.9% and 78.6% of variation in the sensory and objective TPA data, respectively. The PCA biplots appeared to group the cultivars on a similar basis, and results presented here for PCA and PLS regression suggest that objective measurements such as objective TPA values and/or DM data alone could be used to characterise some aspects of the sensory texture and flavour of squash.