A total of 14 seedbome fungi were identified among the cotton (Gossypium
barbadense L.) cultivars Giza 80, Giza 83, Giza 85, Giza 86, Giza 88, and Giza 89. No
single cultivar yielded all the 14 fungi. Giza 88 and Giza 89 yielded the highest number
of fungi (12 fungi), while Giza 86 yielded the lowest number (8 fungi). The other cultivars
yielded a number of fungi ranged from 9 to 11. Penicillium spp., Altemaria altemata,
Aspergillus flavus, Aspergillus niger, Cephalosporium sp., and Fusarium moniliforme
were isolated from all the tested cultivars. Cluster analysis of the cultivars, based on
their fungal profiles, showed that Giza 80 had a unique fungal profile different from the
profiles of the other cultivars. On the other hand, Giza 86 and Giza 89 showed a very
high similarity level in their fungal profiles. Proteins of the cultivars were separated by
SDS-PAGE, and the obtained banding patterns were visualized by using Coomassie-
Blue staining system. Data for frequencies of the isolated fungi (dependent variables)
and amounts of protein fractions (independent variables or predictors) were entered into
a computerized stepwise multiple regression analysis. Using the predictors supplied by
stepwise regreSSion, 12 regression models were constructed to predict frequencies of
the isolated fungi. Coefficient of determination (R2) values of the models ranged from
62.72 to 100%. These results suggest that SDS-PAGE of proteins may provide a
supplementary assay to microscopic examination to quantify fungal profiles of cotton
seeds.