Fig. 2: Two rows of figures at top plot the Resnik-based similarity vs. NAF between mouse-human (top row) and yeast-human (bottom row).
From: SANA: cross-species prediction of Gene Ontology GO annotations via topological network alignment
Top Table: Pearson correlation (ρ) and statistical significance of the plots. Middle Table: Filtering for well-annotated proteins, we see higher Pearson correlations between NAF and Resnik score (allowing all evidence codes) that result when filtering for well-annotated protein pairs in EC-driven alignments; N is the number of aligned protein pairs for which both proteins are annotated with at least S GO terms that are each annotate at most M proteins per species. We exhaustively list every pair of BioGRID species for which the Pearson p-value is <10−2 for S ≥ 8 and M ≤ 4; the table is sorted by ρ(S, M). Bottom Table: Pearson correlation between M, S, and ρ(S, M) above, across all species and values M and S for which ρ(S, M) was statistically significant (see text for further discussion).
