Figure 2.

Examining different metrics to predict GO categories for a gene using a set of its co-expressed genes. Addition of Pearson’s correlation coefficient to calculation of co-expression score (equation 1, see also [5]) increased the number of predicted ontologies equal to canonical GOs previously assigned to a gene (number of direct hits, (red vs. light blue lines). Using top 20 co-expressed genes instead of top 40 decreased the number of direct hits (red vs. dark blue lines).

Dozmorov et al. BMC Bioinformatics 2011 12(Suppl 10):S14   doi:10.1186/1471-2105-12-S10-S14