Exons were identified after vector screening using transcript, rodent, and protein databases. The definition of a record varies according to the database, while 'exons' refer to high-scoring segment pairs in BlastN comparisons (E < 10-15 and sequence identity >90%) to the genome. Unique exons and all subsequent columns refer to placements that were possible after considering the preceding databases. Placement of rodent transcripts required evidence of splicing and sequence identity >80%. ORFs were identified using getorf [84] using a minimum size of 30bp to report. Protein homology required BlastX E < 10-15. Pfam hits required score >20 using hmmpfam [92]. Gene prediction programs are described in Table 2. CpG islands were identified using cpgreport [84] using standard criteria [45].