Figure 3

Sequence of clustering dendrograms from 27K blood reference data set. The dendrograms display the results of clustering the non-intercept columns of the coefficient matrices Δ _k by applying Euclidean metric and Ward’s linkage method. Note that the intercept (reference) represented whole blood. The sequence shows that differing the values of k results in distinct levels of information with respect to cell lineage. Red dots indicate myeloid lineage (granulocyte and monocyte).