Economically, the Rosaceae is one of the most important plant families  comprising some 90 genera with over 3000 distinct species having chromosome numbers ranging from x = 7 to x = 17 . Four sub-families are distinguished on the basis of fruit types: the Maloideae (including Malus, and Pyrus); the Prunoideae (Prunus and other stone fruit and almonds); the Rosoideae (Fragaria, Rubus, and Rosa); and the Spiraeoideae (containing many ornamental species, including Physocarpus) . A recent phylogenetic treatment of the Rosaceae based on DNA sequence data of nuclear and chloroplast genomic regions reclassified the genus into three sub-families (the Dryadoideae, the Rosoideae and the Spiraeoideae), each containing a number of distinct supertribes . Prunus and Malus are included in the Spiraeoideae, supertribe Amygdaleae and Pyrodae (tribe Pyrinae) respectively, whilst Fragaria is included in the Rosoideae, supertribe Rosodae (tribe Fragariinae). It has been postulated that the poor phylogenetic resolution along the backbone of the Rosaceae phylogenetic tree suggests a rapid evolutionary radiation of lineages within the family, corresponding to a relatively recent divergence of the genera . Possibly because of the rapid evolution, members of the Rosaceae display remarkable phenotypic diversity, with common morphological synapomorphies not readily identifiable. Indeed, plant habit, chromosome number, and fruit type have all evolved independently on more than one occasion within the family [2, 4, 5]. A better understanding of how the phenotypic diversity within the Rosaceae arose would provide an insight into how evolution can lead rapidly to diversification.
Comparative mapping has been carried out in a number of economically important plant families including the Poaceae, Solanaceae, Brassicaceae and Fabaceae [6–10]. In the Poaceae, marker order is highly conserved within syntenic 'genome blocks' between genera . However, despite the conservation of syntenic blocks, grass lineages may rapidly evolve, with high rates of chromosomal 'reshuffling' observed between the rye and wheat genomes . Using restriction fragment length polymorphisms (RFLPs), Bonierbale et al.  studied the conservation of synteny between potato (Solanum tuberosum) and tomato (S. lycopersicum syn. Lycopersicon esculentum) and found remarkable conservation of genome structure, with only a few regions where paracentric chromosomal rearrangements could be identified. More recently, Wu and Tanksley  reported a higher frequency of inversions than translocations among the genomes of different genera of the Solanaceae.
In the Brassicaceae, almost complete genome colinearity between Arabidopsis thaliana and Capsella rubella has been observed, with gene repertoire, order and orientation highly conserved  and likewise, soybean linkage group A2 was shown to be conserved over its entire length with Arabidopsis chromosome I, with just 3 rearrangements identified between the chromosomes of the two species . However, between Arabidopsis and Brassica oleracea, rates of chromosomal rearrangements were shown to be much higher . Between dicotyledenous families, comparisons have been performed between much wider evolutionary distances, for example between Prunus and Arabidopsis [17–19], but only fragmentary patterns of conserved synteny have been observed. However, intrafamilial studies have shown that genome evolution within a family usually proceeds through whole-scale inversions and translocations between chromosomes, meaning regions in which marker order is highly conserved can be identified between genera that diverged millions of years ago, and thus information on genes within conserved genome blocks of one genus can inform studies in other genera within a family.
Comparative genome studies in the Rosaceae have so far been based on the alignment of genetic maps within the same genus, or amongst closely related genera, using small sets of orthologous markers. These studies showed that the genomes of Prunus species are essentially collinear, for example in peach and apricot [20, 21], and peach and sweet cherry [22, 23]. Similarly, within the Pyrinae tribe, the genomes of Malus, Pyrus and Eriobotrya were shown to be highly collinear [24–26]. Only a few studies investigated genome comparisons across Rosaceae tribes or subfamilies. Dirlewanger et al.  compared Malus and Prunus and found strong evidence that single linkage groups in the diploid Prunus were homologous to two distinct homeologous linkage groups in the amphitetraploid genome of Malus. Vilanova et al.  compared the diploid reference linkage maps for Prunus (T×E; almond 'Texas' × peach 'Earlygold') and Fragaria (FV×FN; F. vesca ' 815' × F. nubicola '601') and they identified numerous chromosomal translocations and rearrangements that occurred in the 29 million years since the genera diverged from a common ancestor. They also found clear cases of conservation of chromosomal synteny, and reconstructed a hypothetical ancestral Rosaceae genome composed of nine chromosomes.
Whole genome sequencing using next generation technologies has now become accessible to the broad scientific community. In the Rosaceae, the genome of Malus × domestica was recently sequenced using a whole genome shotgun approach . The analysis of the draft sequence of 'Golden Delicious' is consistent with a putative nine chromosome diploid ancestor for the genus . Although there are no published genome sequences that would permit direct comparisons of the M. × domestica genome and those of other genera of the Rosaceae, there are a large number of genetic markers available for Rosaceous species. Recently, Cabrera et al.  reported the development of 857 Rosaceous Conserved Orthologous Set (RosCOS), of which 613 were mapped on the T×E Prunus reference map. Because of the conserved nucleotide sequence and low or single copy presence across species, the COS sequences are particularly useful for comparative genome studies between related species . The RosCOS set demonstrated extensive conservation of synteny between poplar and Prunus, members of two different plant families in the eurosid clade . Also a significant fraction of the RosCOS markers were transferable to the FV×FN Fragaria reference population.
In this paper, we compared the apple genome sequence with conserved molecular markers previously mapped in Prunus and Fragaria [28, 30, 31], along with an additional set of markers developed for the purpose. The goal was to derive at the ancestral genome structure and organisation of species within the Rosaceae family and to study the genome evolution of the economically relevant Fragaria, Malus and Prunus. The locations of molecular markers on the Prunus and Fragaria genomes were determined through bin mapping [32, 33], and their physical positions through analysis of the Malus × domestica ' Golden Delicious' genome . The comparisons identified syntenic blocks common among the genomes of the three genera as well as within the polyploid Malus genome. The findings allowed us to hypothesize about genome evolution within Rosaceae, and to reconstruct its ancestral genome for the family.