The map presented here is the first map of all 18 porcine autosomes based on gene-associated SNPs and it contains 330 not previously located genes. The positions of 81 SNP-containing sequences representing these genes were confirmed by Blast analysis against the newly assembled BAC porcine sequence (Additional file 1
A total of 456 gene-associated and six porcine EST-based SNPs identified from re-sequenced exons were mapped to the 18 porcine autosomes. The main selection of the SNPs was unbiased as they were found randomly in the amplicons containing the EST sequences, which were distributed across the 22 human autosomes. Almost all SNPs showed heterozygosity in the sires except 14 SNPs that were included because of interest to other projects. As the SNPs were selected on linkage ability with a very high two-point LOD score (>55) the number of informative meioses is expected to be high. In the present study it ranges from 401 to 6,898 allowing us to calculate a very robust map. The family comprised Duroc sires crossed with Danish Landrace/Danish Large White sows and as the SNPs were selected from alignments of re-sequenced exons in the 12 Duroc sires this implies that most of the informative meioses arose from this population. For the sow population we considered instead the minor allele frequency (MAF) of each SNP which gave us an idea of the distribution between male and female meiosis. The analysis did not indicate that the difference in meiosis numbers affected the results as the number generally was very high (Additional file 2
The length of the averaged-sex linkage map was calculated to be 1,711.8 cM, the female map was 2,336.1 cM and the male map was 1,441.5 cM. As described in the result section when comparing to the map of Rohrer and colleagues [2
] our map is about 20% shorter. The animal material they used to create their linkage map is much smaller than the material used here, which might influence the recombination rates at the ends of the chromosomes and thereby overestimate the length. We see this phenomenon on the female map when the number of meioses is low like on SSC 17p where the MCM8 marker is positioned 100 cM from MKKS (Figure ). The first whole-genome map estimated the female map to be about 21 Morgan (M) and the male map around 16.5 M [1
]. When these female and male maps are compared a recombination ratio of 1.3:1 is found. Another estimation of the ratio suggests the recombination ratio to be 1.55:1 [3
]. In our case the ratio was even higher, approximately 1.65:1, resembling the human recombination ratio of 1.7:1 [21
When considering each chromosome map of the female and male only three chromosomes separate from the rest (Figure ). The SSC 11 and the SSC 13 differ from the rest by having a longer male map. Regarding SSC 11, this is in accordance with previous work where a longer p-arm of the SSC 11 male map was found [22
]. However, these authors also reported that the complete male map of SSC 11 was shorter than the female map, which could be the case in the present study too if the entire map of SSC 11 was available. On SSC 1 there is a clear difference in how the distance of the markers on the female and male maps varies across the entire chromosome as reported previously [23
]. However, the female map presented here is longer than the male map, which is due to a single marker (C9orf78) positioned at the telomeric end of SSC 1q. The MAF found in the sow population of this marker is low leading to few female informative meioses, which can explain the high recombination rate between the markers NDUFA8 and C9orf78 on the female map.
The linkage analysis showed that most synteny groups were present and only few single gene and micro rearrangements were found. Three synteny regions at SSC 1 were not represented on the linkage map, i.e. the HSA 14q21 region and both extremes of HSA 18. The HSA 22q12-qter, which mapped to SSC 5 [19
], was also not present on our map. In addition to the absence of these regions, segments from four human chromosomes were missing on SSC 2 (HSA 1), SSC3 (HSA 9) and SSC 17 (HSA 4 and 8) when comparing to the comparative segments identified by Meyers and colleagues [11
]. These regions were missing due to the fact that no SNPs were identified in exons representing these genomic areas. However, on SSC 3, 7, 8, 9, 10, 15 and 16 the map presented here contains nine additional regions. In addition, the telomeric ends of SSC 3 and 4 were divided in more segments than the previous comparative map [11
], though in general the present map is divided in fewer segments, probably due to fewer markers. Apart from this, the similarity concerning conserved synteny between these two maps is very high.
All single gene rearrangements and SNPs denoting porcine ESTs were analysed to verify the linkage order. A total of ten genes and six ESTs were mapped onto the IMpRH7000 panel and in all cases the location was confirmed (Figures , , , and ). The six ESTs were all mapped to the human genomic sequence. In three cases the location matched the expected synteny group but for P0497, P0150 and P0337 the human homologous region could not be determined.
The nine new regions discussed above were mapped onto the IMpRH7000
panel to confirm their locations. These regions refer to the following genes: PCBD1
from HSA 10q22.1 on SSC 3; NDRG3
from HSA 20q11.23 and HSA 17q12, respectively, on SSC 7; OSTF1
from HSA 9q21.3 on SSC 8; MYLIP
from 6p22.3 on SSC 9; LIN37
from HSA 19p13.12 on SSC 10; YWHAB
from HSA 20q13.12 on SSC 15 and finally IDH1
from HSA 2q34 and HSA 19p13.2, respectively, on SSC 16. None of these genes/human regions have previously been mapped to these porcine chromosomal locations [11
], but NDRG3
were confirmed by Blast of the SNP-containing sequence against the recently assembled porcine map [13
]. Moreover, all gene locations and orders in relation to other markers were confirmed by linkage to the microsatellite RH-map [20
]. Most of the single gene rearrangements were located in regions between larger synteny groups, which also support the calculated marker order on this genetic map.
Of the 440 genes and ESTs located on the map, a total of 110 genes have previously been mapped either by linkage or physically and all of these except one were mapped to the expected chromosome. The F13B
gene located at HSA 1q31.3 deviated as it was formerly mapped by linkage analysis to SSC 4 [24
], but as this fragment of the human genome is known to correspond to a region on SSC 10 [10
] the difference is highly likely. The linkage result of F13B
to SSC 10 was confirmed by mapping the gene onto the IMpRH7000
Only two previously physically mapped genes were ordered differently than expected. The first was the TPM4
, which was mapped to SSC 2q24-q29 [25
]. In the present study it was positioned further up on the linkage map than MEF2C
on SSC 2q21-q22 (Figure ). On the comparative map of this region, only TPM4
from HSA 19 is located among genes from HSA 5 [19
] which indicates that our positioning of the TPM4
gene among other genes from HSA 19 might be correct. The other gene that was located at a different position than expected was FUT1
from SSC 6q11 (Figure ). This gene was previously cytogenetically mapped closer to the centromeric region than the genes from SSC 6q12 [26
]. In our analysis the gene was located more distal, close to the XRCC1
gene from the same human chromosomal region on HSA 19q13.31. The XRCC1
gene has previously been mapped to SSC 6q12-22 [27
] which might be the actual position of the FUT1
gene as well.
A total of 81 genes were matched by Blast analysis of SNP-containing sequences against the PreEnsembl porcine BAC sequence. For nine short regions containing one to three markers on the chromosomes, minor rearrangements of the gene order occurred (marked in bold in Additional file 1
). Within these regions of minor rearrangements the distance between the markers was small both on the linkage map and on the BAC assembly. The discrepancies could be caused by rearrangements in the BAC assembly of the porcine chromosomes or deviations on the genetic map, due to missing segregations in some boar families despite the SNP network resulting from sows having litters with more than one boar. However, these areas should be subject to further investigation.
This SNP map forms the basis for ongoing studies on QTLs for meat quality, growth, osteochondrosis, lung diseases and other traits. An advantage of the map is the ability to compare to previously reported QTL studies made on microsatellite markers as well as the fact that 138 of the SNPs is represented on the new Illumina PorcineSNP60 Genotyping BeadChip [18
]. The actual SNP overlap can be identified through the PorcineSNP60 name (Additional File 2
). Therefore, the map can function as an anchoring map for future maps created by use of Bead Chip technology. Furthermore, new SNPs can be added to this version of the map for fine-mapping.