Distribution of Rule Types
| Rule groupa | Total number of rulesb | One similar GO proteinc | Single functionc | Consensus leafc | Near consensus leafc | Consensus ancestorc |
| ProDom-All | 17504 | 11498 (66%) | 2466 (14%) | 1251 (7%) | 472 (3%) | 1817 (10%) |
| ProDom-NoIEA | 10724 | 7579 (71%) | 1361 (13%) | 540 (5%) | 176 (2%) | 1068 (10%) |
| CDD-All | 1792 | 651 (36%) | 337 (19%) | 179 (10%) | 161 (9%) | 464 (26%) |
| CDD-NoIEA | 1396 | 639 (46%) | 249 (18%) | 105 (7%) | 84 (6%) | 319 (23%) |
[i] The number and percentage of rules and rule-types for each combination of domain database (ProDom or CDD) and set of protein GO associations (All or No IEA) are shown. No “near consensus ancestor” rules were ever created. aProtein domain database and types of GO associations considered when building rules.bTotal number of rules built. cNumber and percent of rules as a function of rule types illustrated in Figure 3.