Additional File 1:

Global properties of the probability assignment schemes. Shows properties like average and median probabilities.

Format: DOC Size: 30KB Download file

This file can be viewed with: Microsoft Word Viewer

Additional File 2:

Correlation of interaction probabilities with mRNA expression correlation. Ribosomal components are among the most co-expressed genes, and could potentially lead to the observed relative importance of co-expression data. To check for the effect of ribosomal proteins, we filtered the yeast interaction set in our analysis to remove all ribosomal proteins and calculated the correlation between co-expression and interaction probability. These results are shown in Additional Table 2. The numbers in the brackets represent the values of Spearman correlation coefficient and weighted average after removing the ribosomal proteins from the interaction data. We observe that removing the ribosomal proteins does not change the values significantly.

Format: DOC Size: 28KB Download file

This file can be viewed with: Microsoft Word Viewer

Additional File 3:

Histograms of GO similarity scores. We evaluated the GO similarity scores for known yeast interactions reported in the MIPS database [26]. The histogram of the scores is shown in Additional Figure 1A. We also generated a background distribution by computing the GO similarity scores for 1,000 random interactions (Additional Figure 1B). These random interactions were generated by picking pairs of proteins randomly from the set of interacting proteins in yeast. It is evident from the two figures that true proteins interactions (i.e known yeast interactions reported in MIPS) generally have lower GO similarity scores than the background.

Format: DOC Size: 82KB Download file

This file can be viewed with: Microsoft Word Viewer

Additional File 4:

Spearman partial correlations for schemes using expression as input. Spearman Partial Rank Correlation Coefficient. The Spearman partial rank correlation coefficient between two random variables A and X, given the fact that both A and X are correlated to random variable Y, denotes the correlation between A and X, when Y is kept constant. It is calculated as follows:

Here, r_AX, r_XY and r_AY represent the Spearman correlation coefficients between A and X, X and Y, and, A and Y respectively. The significance level is given by

D_{AX, Y} has a normal distribution with zero mean and variance one. N represents the size of the data set.

Format: DOC Size: 25KB Download file

This file can be viewed with: Microsoft Word Viewer