Table 2.
Biological meaning of WordCluster predictions*
| Method | #islands | #TSS overlap | #R13 overlap | #Alu overlap | #PhastCons overlap |
|---|---|---|---|---|---|
| cpg50 | 198703 | 12432 (6.3%) | 30660 (15.4%) | 80323 (40.4%) | 48787 (24.6%) |
| cpgISc | 194724 | 11926 (6.1%) | 34567 (17.8%) | 70144 (36.0%) | 48930 (25.1%) |
| cpgISg | 204238 | 12156 (6.0%) | 37616 (18.4%) | 70456 (34.5%) | 52335 (25.6%) |
*Comparison of three WordCluster predictions of CG clusters (CpG islands) using three different distance models: cpgISg (genome intersection), cpg50 (median) and cpgISc (chromosome intersection). The overlap with two gene regions (TSS and R13), Alu elements and phylogenetically conserved PhastCons elements have been measured and both absolute numbers and percentages are given.