Isozyme comparisons using weighted analyses of electrophoretic mobility in a range of polyacrylamide gel concentrations

Peroxidase isozymes from two flax genotypes were separated electrophoretically using a range of concentration, regression models were fitted by weighted least squares method, using matrix techniques and appropriate programmes, weight were reciprocals of variances of significance, as a chi square, of the residual sums of square (SS). Log (Rm)Plotted against gel concentration showed significant curvilinearity; Log (P/Hb), where P was an isozyme's migration was linear with concentration. Matrix method allowed a completely generalized approach to the analysis of data of this type. In particular, regression data could be readily combined so as to allow orthogonal comparison between Intercepts and slops in an analohgous fashion to the orthogonal breakdown of row, column and interaction effects in a balance two-way analysis of Weighting effected some improvement in comparison between isozymes; combined analyses of weighted regression produced striking improvement in discrimination relative to simple pairwise Comparisons between isozymes.