Figure 4.

Classes of overlapping annotated ORFs in S. cerevisiae derived from comparison with the A. gossypii genome. Class A, homologs for both overlapping S. cerevisiae ORFs are found at syntenic positions in A. gossypii and also overlap. Class B, homologs for both overlapping S. cerevisiae ORFs are found in A. gossypii but do not overlap (see comments at end of legend). Class C, gene X but not gene Y has a syntenic homolog in A. gossypii. Class D, both overlapping S. cerevisiae ORFs lack a homolog in A. gossypii. Numbers refer to the frequency of the four types of overlapping pairs. Although all three possible directions of overlaps were observed, for convenience only 3'/3'-end overlaps are depicted here. YPL166W/YPL165C and YLR360W/YLR361C are the only two cases for which overlap was observed both in S. cerevisiae and A. gossypii (class A). The two overlaps are short (24 and 35 nucleotides, respectively, in S. cerevisiae and 24 and 14 nucleotides in A. gossypii, and involve only terminator-ORF sequence overlap. Class B overlaps comprise three syntenic ORF pairs (YJR012C/YJR013W, YML095C/YML096W, and YGR074W/YGR075C) and two non-syntenic (YPL018W/YPL017C and YBR262C/YBR263W). The lack of synteny reflects chromosomal rearrangements in one or the other species which resulted in either the separation of the two ORFs in A. gossypii or in their joining in S. cerevisiae. Two of the five class B ORF pairs refer to YML096W/YML095C and YGR074W/YGR075C, both with 6-nucleotide 3-end overlaps. The syntenic A. gossypii homologs are separated at their 3' ends by 2 and 51 nucleotides respectively, implying some overlap of terminator and ORF sequences very similar to their syntenic S. cerevisiae homologs. For two pairs of ORFs overlapping at their 5' ends in S. cerevisiae (YBR262C/YBR263W and YJR012C/YJR013W) both ORFs have a homolog in A. gossypii. The alignments of YBR263W and YJR012C with their respective A. gossypii homologs strongly suggest an error in selection of their start codons. Both S. cerevisiae ORFs are very likely to be 75 codons shorter, thus eliminating the presumptive promoter-ORF overlaps. Directions are provided for classes C and D in the additional data files.

Brachat et al. Genome Biology 2003 4:R45   doi:10.1186/gb-2003-4-7-r45