Table 2.

Construction of K_5quali with values for the protein similarities based on the information about missing and available data. Each cell of the table contains the similarity between two proteins x and x^′. The values to complete missing data were ε=0.01, θ=0.02, the mean similarity of the detected protein x∊X₁ with other proteins inside X₁ (m_x) and the overall mean similarity for the comparison of two nonobservable proteins (X₃). As the similarity between the protein and itself is maximal, the diagonal of K_5quali is composed of ones.

	Observed X1	ObservableX2	Nonobservable X3
Observed X1	$K_5^{*}(x,x^{\prime })$	ε	$m_x=\sum _{\begin{array}{c}{x}^{″}\ne x\\ x^{″}∊X_1\end{array}}\frac{K_5^{*}(x,x^{″})}{[\mid X_1\mid -1]}$
Observable X2	ε	θ>ε	ε
Nonobservable X3	$m_x=\sum _{\begin{array}{c}{x}^{″}\ne x\\ x^{″}∊X_1\end{array}}\frac{K_5^{*}(x,x^{″})}{[\mid X_1\mid -1]}$	ε	$\frac{1}{\mid X_1\mid }\sum _{x∊X_1}{m}_x=\overline{m}$