We used a strategy based on the comparative mapping of sequence contigs from the chrUn fraction of the chicken genome and from chicken ESTs to characterize a chicken microchromosome that was absent from the first sequence assembly. Isolation of BAC clones and FISH mapping enabled us to identify this microchromosome as being GGA25. Although GGA25 is now included in the second version (May 2006) of the chicken genome assembly, our RH maps enabled us to assign 23 additional contigs from the chrUn, amounting to a total of 44.8 kb of sequence, and 18 ESTs that did not have any significant BLAST hit. Genetic maps of GGA25 were successfully developed by using the data from the RH maps.
The difficulty we encountered in isolating BAC clones and the small size of many of the chrUn contigs we attributed to GGA25 suggest that there may be a cloning and/or a sequencing bias against obtaining sequence from microchromosomes. These difficulties in obtaining sequence have since been confirmed for other microchromosomes by us and others (J. Dodgson, personal communication). The GGA25 sequence in the May 2006 assembly has a composition of 52.4% G+C, which is amongst the highest for chicken chromosomes . The chrUn contigs that we have added to GGA25 through RH mapping have a mean G+C content of 61.2% [see Additional file 1]. This confirms GGA25 to be amongst the most GC-rich microchromosomes. The observation from FISH experiments , and the high proportion of minisatellite markers on the E26C13 genetic map , suggest that GGA25 contains a high number of repetitive sequences. Both the high G+C content and a high proportion of repetitive elements may account for the paucity of available sequence for this chromosome in the chicken genome assembly at the beginning of this study.
An average retention frequency of 24.1% (range 11.2–39.3%) was observed for the 120 GGA25 markers studied here. This finding is in agreement with previously observed values [18, 19], confirming a higher retention rate for micro - than for macrochromosomes.
By using the average value of 38.7 kb/cR observed for other chromosomes mapped with the ChickRH6 panel , the length of GGA25 can be estimated to be around 11.4 Mb, which is in accordance with the expected size for this chromosome . This suggests also that the entire chromosome is covered by our RH map. The portion of human chromosome 1 orthologous to GGA25, located at HSA01q21.2-q23.3, is about 11.5 Mb long (from 148.1 to 159.5 Mb, in addition to a short fragment of about 100 kb around 144.2 Mb, Hg18). This region corresponds to a chicken microchromosome covering about 11–12 Mb, giving a ratio of about 1 Mb HSA/1 Mb GGA, which is much lower than the average ratio when considering the whole genome (2.4 Mb HSA/1.0 Mb GGA), and may be associated with the very high density of genes in this portion of the human genome [20, 21].
The HSA1 region orthologous to GGA25 contains around 410 genes . Assuming most of these genes are present in GGA25, the gene density of this microchromosome would be 5.9 genes/100 kb, which is comparable to the value of 5.5 obtained for GGA32, and higher than the ratio observed for longer chicken chromosomes, as expected .
The RH results indicate a high number of chromosomal rearrangements in the chicken and human lineages in the region corresponding to GGA25 (Figure 4). The high number of intra-chromosomal rearrangements within the region of conserved synteny between birds and mammals is in accordance with results obtained for other chromosomes, e.g., GGA02 , GGA05 , GGA07 , GGA10 , GGA14 , GGA15  or GGA28 .
Genotyping the East Lansing population allowed us to connect our results to the international chicken reference backcross mapping pedigree, and to develop a single-locus marker from a minisatellite previously mapped in this population. These results suggest that minisatellite-type sequences are distributed throughout GGA25 (Figure 5).
The genetic map constructed using our local cross presents smaller genetic distances between markers than the East Lansing map. The variation observed are most likely the result of differences between the lines used for the two crosses. By using the genetic map built with our local cross, we find an overall ratio of 3.0 cR/cM, which is a relatively low value when compared to the ratio obtained for chicken macrochromosomes [18, 23, 24], but close to the ratio of 3.6 obtained for chromosome 14 . This result is in accordance with the fact that the recombination rate is negatively correlated with the physical length of the chromosome .