Development of predictive models for end-use quality of spring wheats through canonical analysis

被引:42
作者
Butt, MS
Anjum, FM [1 ]
van Zuilichem, DJ
Shaheen, M
机构
[1] Univ Agr, Dept Food Technol, Faisalabad, Pakistan
[2] Wageningen Agr Univ, Dept Food Sci, HTST Grp, Wageningen, Netherlands
关键词
canonical correlations; farinograms; physico-chemical; protein composition; stepwise regression;
D O I
10.1046/j.1365-2621.2001.00477.x
中图分类号
TS2 [食品工业];
学科分类号
0832 ;
摘要
For 30 spring wheat cultivars, the bread-loaf volume, the total bread-baking scores and the total chapati scores were considered as dependent variables, and 16 other quality attributes were regarded as independent variables. A stepwise regression indicated acid-soluble glutenin, globulin, gliadin, and Mixograph development time contributed negatively to the bread-loaf volume and the total bread-baking scores, whereas dough development time, measured by farinograph readings, and particle size index contributed positively to both these attributes. However, 0.05 M NaOH soluble glutenin, acid soluble glutenin, globulin, albumin, fat and test weight positively affected the total chapati scores while the Farinograph readings (dough development time), ash, moisture and falling number had a negative affect. The canonical correlations were computed between the two groups of variables. The first group comprised dependent variables (bread-loaf volume, bread scores and chapati scores) and the second comprised 21 independent variables. The highest observed value of coefficient of determination for white flour protein was 0.8912 indicating that 89.12% of the variation in white flour protein can be explained by all other variables in the second set. However, 24.36 and 24.60% of variation in bread-loaf volume and total bread scores respectively can be explained by all other variables in the first set. The squared multiple regression of each variable in the first set with all the variables in the second set indicated that 57.13% variation in total chapati scores was explained by the variable in the second set, whereas 50.74% of the variation in the bread-loaf volume and 34% of the variation in total bread scores was explained by the variables included in the second set. The results of the canonical correlation revealed high correlation between the end use quality characteristics and other physico-chemical, rheological and protein components of wheat cultivars.
引用
收藏
页码:433 / 440
页数:8
相关论文
共 12 条
[1]  
AACC, 1983, Approved methods of the American Association of Cereal Chemists, V8th
[2]  
ANJUM FM, 1976, THESIS U AGR FAISALA
[3]  
DIXON WJ, 1990, BMDP STAT SOFTWARE M, V2
[4]  
GITTENS R, 1985, CANONICAL ANAL REV A
[5]  
GRAYBOSCH R, 1993, CEREAL CHEM, V70, P95
[6]  
Hotelling H, 1936, BIOMETRIKA, V28, P321, DOI 10.2307/2333955
[7]  
LUKOW OM, 1990, CEREAL CHEM, V67, P170
[8]  
MUHAMMAD F, 1995, STAT METHODS DATA AN
[9]  
NG PKW, 1988, CEREAL CHEM, V65, P408
[10]  
Pomeranz Y, 1968, Adv Food Res, V16, P335, DOI 10.1016/S0065-2628(08)60361-4