Introduction
Numerous hereditary diseases of the horse have been described, including rhabdomyolysis, laryngeal hemiplegia, insect bite dermal hypersensitivity, chronic obstructive disease, polysaccharide storage myopathy and osteochondrosis. Many of these are of a quantitative genetic nature. Quantitatively inherited traits show either continuous or discontinuous variation (the absence or presence of a disease). However, the influences of many genes and the environment combine to contribute to a particular phenotype of a quantitatively inherited trait.
With the availability of dense, highly informative marker maps for different species, it is possible to identify DNA segments which cosegregate with the trait in question and to carry out positional cloning of genes. The recent publication of the first comprehensive low-density linkage map (SWINBURNE et al. 2000) and of a second low-density, male-based linkage map (GUERIN et al. 2003) of the horse with markers assigned to every chromosome provides the chance to map traits in the horse by whole genome scanning. The aim of the present study was to establish a highly polymorphic microsatellite marker set covering the whole equine genome. Multiplex sets were developed in order to make the performance of whole genome scans more efficient.
Material and methods Markers
A total of 157 microsatellite markers was chosen from the linkage map generated by SWINBURNE et al. (2000), and from the HORSEMAP database at the INRA Biotechnology Laboratories Home Page (http://locus.jouy.inra.fr), which presently contains 735 microsatellite markers.
Markers were first selected according to their location on the equine linkage map to provide a coverage of the entire equine genome as complete as possible, and then selected according to their possible information content for linkage analysis. The threshold was set at five alleles
Characterisation of a microsatellite marker set for genome-wide screens of the equine genome 28
and with a heterozygosity and polymorphism information content (PIC) greater than 50%. In the published linkage maps or studies, not all markers were provided with number of alleles, heterozygosity or PIC. The number of alleles was not given for four (2.5%) of the markers selected for this work. According to previous reports six (3.8%) had three alleles, 24 (15.3%) had four alleles, and the remaining 123 (78.3%) markers had a number of alleles n ≥ 5. The heterozygosity value was not given for 18 (11.5%) markers and a heterozygosity level
< 50.0% was reported for 17 markers. Heterozygosity values > 50.0% were found in literature for the majority of the 157 markers (122 = 77.7%). Markers with missing values for the number of alleles and heterozygosity were nonetheless chosen for to their position on the linkage map. The majority of the selected markers were dinucleotide repeats. All chosen markers were equally spread on all 31 equine autosomes (ECA) and the X chromosome (Appendix 1, Figures 1 - 4), and the average distance between the markers was not to exceed 20 cM.
Genotyping of microsatellites
All equipment, materials, and chemicals used for DNA isolation, polymerase chain reaction (PCR) and gel electrophoresis are listed in Appendix 3.
The marker set was tested on a reference group consisting of 86 Hanoverian Warmblood Horses, including 46 (53.5%) females and 40 (46.5%) males. Using the QIAamp® 96 DNA Blood Kit (QIAGEN, Hilden, Germany)1.35 µg genomic DNA was isolated from 75 µl EDTA blood.
The PCR amplification of the 157 markers was conducted using PTC 100™ or PTC 200™
thermocyclers (MJ Research, Watertown, MA, USA) and a general PCR programme with variable annealing temperature (Ta) (Appendix 2, Table 10). The reaction started at 94 °C for 4 min, followed by 35 cycles at 94 °C for 30 s, Ta for 1 min, 72 °C for 30 s, and at 4 °C for 10 min.
The PCRs were performed in 15-µl volumes using 10 ng DNA, 10x incubation buffer containing 15 mM MgCl2, 5% DMSO, 100µM each dNTP and 1.0 U Taq Polymerase (Qbiogene, Heidelberg, Germany). The amount of primers, which ranged between 1.0 to 18.0 pmol, was adjusted empirically in order to equalise the signal for each marker in the multiplex groups. All forward primers were flourescently labelled at the 5’ end with IRD700.
Characterisation of a microsatellite marker set for genome-wide screens of the equine genome 29
Ninety-four primer pairs were pooled into PCR multiplex groups of two pairs to increase efficiency, and the 63 remaining primer pairs were PCR amplified separately.
For the examination of the marker genotypes, the PCR products were size fractionated by gel electrophoresis on an automated sequencer (LI-COR 4200/S-2, Lincoln, NE, USA) using 6%
polyacrylamide denaturing gels (rotiphorese®Gel 40, Carl Roth GmbH & Co, Karlsruhe, Germany). Prior to loading, dilution of PCR products with formamide loading buffer was determined empirically and carried out accordingly. To ensure adequate size fractionation of alleles, the diluted PCR products were denatured at 94 °C for one minute. The multiplex groups and the separately amplified PCR products were pooled according to their size and labelling prior to electrophoresis. Allele sizes were detected with IRDye700-labelled DNA ladder, and the genotyping data were analysed manually.
Statistical parameters for the derived marker set
The relative informativeness of the marker set was evaluated by means of the observed heterozygosity (HO), the expected heterozygosity (HE), and the polymorphism information content (PIC) (Appendix 2, Table 10). Only female animals were involved in the calculation for the X chromosome.
The observed heterozygosity indicates the proportion of heterozygote individuals in regard to the number of genotyped individuals in a given population. For example, if 50 of 100 genotyped animals are heterozygous, the observed heterozygosity is 50% or 0.50.
The expected heterozygosity depends on the number and the frequency of the alleles of a marker. The higher the number of alleles and the lower the frequency of the most frequent allele, the higher is the expected heterozygosity. The HE value of a locus with n marker alleles can be calculated as:
where n is the total number of alleles in the population and pi is the population frequency of the i-th allele
The polymorphism information content is defined as the probability that the marker genotype of a given offspring will, in the absence of crossing-over, make it possible to deduce which of the two marker alleles of the affected parent the offspring received. The PIC value is thus the most suitable value for describing the usefulness of a marker for linkage analysis.
Characterisation of a microsatellite marker set for genome-wide screens of the equine genome 30
The PIC value of a locus with n marker alleles can be calculated as:
PIC = 1-
∑
where n is the total number of alleles in the population, pi is the population frequency of the i-th allele and pj is the population frequency of the j-th allele.
Results
The characteristics of the whole marker set including 157 microsatellite markers is given in Table 3. The characteristics established for each marker were the number of alleles, observed heterozygosity (HO), expected heterozygosity (HE), and PIC. Furthermore, the distance between the markers is given according to the linkage map of SWINBURNE et al. (2000).
Table 3. Characteristics of the 157 equine microsatellite markers
Characteristic Min Max Mean SD
Number of alleles 2 13 6.1 2.0
Observed heterozygosity 18.6 91.9 65.7 14.6
Expected heterozygosity 19.2 85.2 64.6 13.7
Polymorphism information content 18.4 83.6 60.3 14.1
Distance between the markers (cM) 4 47 19.1 8.3
The average number of alleles was 6.1, with values ranging from 2 (AHT29, UMNe229) to 13 (AHT28, SGCV23). The mean of the distance between the markers was 19.1 cM. The average value of the observed heterozygosity was 65.7%, with a minimum of 18.6% (HTG02) and a maximum of 91.9% (COR017). The mean values of the expected heterozygosity were 64.6%, with a minimum of 19.2% (HTG02) and a maximum of 85.2% (SGCV23). The PIC values ranged from 18.4% (HTG02) to 83.6% (SGCV23), with a mean value of 60.3%.
The distribution of the polymorphism information content of the marker set using classes of 10% is shown in Table 4.
Characterisation of a microsatellite marker set for genome-wide screens of the equine genome 31
Table 4. Distribution of the PIC in 10% classes
PIC (%) Number of markers (n) %
10 – 20 2 1.27
20 – 30 3 1.91
30 – 40 9 5.73
40 – 50 20 12.74
50 – 60 35 22.29
60 – 70 44 28.03
70 – 80 37 23.57
80 – 90 7 4.46
In this study, markers with a polymorphism information content value < 25% were considered to be of little informative content, while those with a PIC between 25% and 50% were classed as moderately informative, and markers with a PIC > 50% as highly informative. Only three (1.91%) markers were of little information content, while 31 markers (19.75%) were moderately informative , and the majority of the 157 markers (n = 123, 78.34%) was highly informative with a PIC > 50%.
Table 5 presents the characteristics of each individual chromosome for the following eight characteristics HE, HO, PIC, average distance between the markers within the linkage groups (LG), range of distance between the markers within the linkage groups, the distance covered within the linkage groups, total distance covered on each chromosome, and the number of markers on each chromosome.
The average distance between the markers within the linkage groups ranged from 11.0 cM (ECA26) to 44 cM on ECA29. No average distances could be determined on ECA07 and ECA28, as these contained only single markers which were not arranged in linkage groups.
On the largest equine chromosome (ECA01) the microsatellite markers covered 162 cM within the linkage groups, but only 11 cM were covered on ECA26. The marker set for the linkage groups on the 31 autosomes and the X chromosome included a total distance of 1685 cM. Assuming conservatively that each linkage group (n = 32) could detect linkage over a flanking area of 10 cM on each side, this marker set covers a total distance of 2325 cM.
Characterisation of a microsatellite marker set for genome-wide screens of the equine genome 32
Table 5. The equine marker set: Marker features per chromosome
Chromosome HE HO PIC
Characterisation of a microsatellite marker set for genome-wide screens of the equine genome 33
The number of markers on the chromosomes ranged from two on ECA07, ECA26, ECA28 and ECA31 to 12 markers on ECA01. ECA26 had the highest average HE value (77.1%), and ECA14 the lowest (50.3%). The highest average HO value was found on ECA06 (76.3%) and the lowest on ECA22 (49.5%). The highest average PIC value was found on ECA26, 73.7%, the lowest on ECA22, 46.7%.
Discussion
The aim of this work was to establish a genome-wide microsatellite marker set for whole genome scanning in the horse. A total of 157 microsatellites placed on all equine autosomes and the X chromosome, were chosen and tested for their suitability.
The marker set can be considered suitable for whole genome scans in horses for the calculated values of observed heterozygosity (x = 65.7%), expected heterozygosity (x = 64.6%) and polymorphism information content (x = 60.3%). As the animals of the reference group consisted mainly of 13 half-sibling groups, a second calculation was carried out in order to test whether the values of HO, HE and PIC had been underestimated. In this test, one foal of each half-sibling group was chosen at random, and these thirteen foals were used for the calculation of HO, HE and PIC for the 157 markers. This procedure was carried out in four independent samples of foals, and the mean value was calculated afterwards. The mean values of the second calculation (HO = 65.4%, HE = 62.9%, PIC = 59.2%) corresponded well with the mean values of the first calculation. Furthermore, the average values of HO, HE and PIC corresponded well with those identified for South German Coldblood horses by MEYER (2003), who calculated a mean HO of 69.6%, a mean HE of 66.5%, and a mean PIC of 62.4%, respectively. This confirms the suitability of the marker set developed here for horse breeds other than Warmblood.
Only three markers included in the present marker set were of low informativeness (PIC < 25%) and may in future be exchanged if highly polymorphic microsatellites are identified in these regions. The majority of the markers (n = 123, 78.34%) was highly informative, with a PIC > 50%.
The marker set introduced in this study represents a useful tool for QTL analyses in the horse.
The Hanoverian Warmblood horse is composed of many breeds including Thoroughbred, Arabian, Anglo-Arabian and different Warmblood breeds, so that this marker set may be also
Characterisation of a microsatellite marker set for genome-wide screens of the equine genome 34
applicable to those breeds. Furthermore, this marker set has already been proved to be applicable to Coldblood horses (MEYER 2003).
Another criterion for the properties of the marker set was equal spacing of the markers over all chromosomes. The distance between the markers was not to exceed 20 cM because otherwise there is less likelihood of detecting a QTL. As the average distance between the markers was 19.1 cM, the predetermined limit of 20 cM was met. However, there are regions on nearly all chromosomes in which the distance exceeds 20 cM, and the largest gap in the set was 47 cM. Due to the fact that the linkage map generated by SWINBURNE et al. (2000) is a low-density map, some regions did not provide a sufficient number of polymorphic markers.
Furthermore, some of the markers were not suitable for QTL analysis because of their low PIC values. The total distance covered with the marker set was 2325 cM. Assuming that a QTL can be detected in the flanking area of 10 cM on each side of a marker, gaps remain in those regions where the distance between two markers exceeds 20 cM. And if the distances of these gaps are added together, the total distance covered with the marker set has to be adjusted to 2081 cM. This value has to be considered as preliminary because it will probably be possible to assign the unlinked markers to linkage groups, so that the distance covered will increase or decrease, depending on the new positions of the markers. A total map length of 2720 cM has been estimated from chiasma counts in horses (LINDGREN et al. 1998), but this length may be higher. According to the length estimated by LINDGREN et al. (1998), the marker set developed here covers 76.5% of the horse genome.
In two recently published independent studies whole genome scanning was carried out with different marker sets in order to assign the horse grey coat colour gene to ECA25 by linkage analysis (SWINBURNE et al. 2003, LOCKE et al. 2003). The number of markers (n = 157) used in our panel is nearly as large as the number (n = 173) used in the study of SWINBURNE et al (2003), and greatly exceeds the number of markers (n = 101) used by LOCKE et al. (2003). However, the three marker sets cannot be compared directly, as not all of the microsatellites chosen in this study are the same as those used by SWINBURNE et al.
(2003) or LOCKE et al. (2003). Furthermore, the values for the observed heterozygosity, expected heterozygosity and PIC of the marker set and the individual markers are not indicated in the other two studies.
At present there are four different known linkage maps: those of LINDGREN et al. (1998);
GUERIN et al. (1999); SWINBURNE et al. (2000), and the recently published second generation half-sibling linkage map by GUERIN et al. (2003). Furthermore, a large number of
Characterisation of a microsatellite marker set for genome-wide screens of the equine genome 35
syntenic, comparative and cytogenetic maps of different single chromosomes have also been published (e.g. CAETANO et al. 1999a, b; GODARD et al. 2000; LEAR et al. 2001;
MARIAT et al. 2001; MILENKOVIC et al. 2002; RAUDSEPP et al. 1996, 1999, 2002;
SHIUE at al. 1999). With the recent generation of the first whole-genome radiation hybrid (RH) map in the horse (CHOWDHARY et al. 2003) an integration of these different maps has been achieved, thus producing the most comprehensive map for the horse at present.
Comparing the positions of the microsatellites in this marker set with the positions of the markers in the RH map of CHOWDHARY et al. (2003), the positions of the markers of 26 chromosomes corresponded well. Sixteen markers used in this study were not located on the RH map, while 23 single markers could be assigned to existing linkage groups. Only on three chromosomes did the positions of markers differ slighty between the two maps. Further newly developed markers will be added to the RH map of CHOWDHARY et al. (2003) in the future.
Another strategy for identifying genes which avoids the necessity of a complete genome scan would be to select candidate genes that have already been shown to cause the same or a very similar disease in other species. The respective orthologous human cDNA or a PCR product of a coding region of this candidate gene can be used for screening the heterologous genomic DNA library in order to isolate genomic DNA clones containing the candidate gene. Once the candidate gene has been mapped, closely localised markers could be identified in the available maps. It is also possible to screen the isolated genomic DNA clones in order to identify new and gene-associated markers. These markers could be used for linkage studies in families segregating for the disease in question. In this way, a combined candidate gene and genome scan approach can be developed for the identification of QTL of equine osteochondrosis.
The primary reasons that any given locus will remain undetected in a genome screen are uninformative markers in the region in critical families, or the pedigrees with insufficient power to provide a statistically significant Lod score at reasonable values of recombination rates (RICHMAN et al. 2001). When using the marker set introduced in this study, attention has to be paid to regions of the genome covered by markers that are uninformative in the selected pedigrees. In such a case, additional markers should be selected from the current most comprehensive map (CHOWDHARY et al. 2003) and tested for informativeness in the parents of the families used in the linkage screen. Generally, particular care should be taken to collect samples from affected members of pedigrees. Large, informative families including both affected and unaffected or non-carrier individuals should be selected for analyses.
Characterisation of a microsatellite marker set for genome-wide screens of the equine genome 36
The marker set for whole genome scans introduced in this study is the first step towards the identification of QTL in the horse. The set can be augmented with additional markers from the RH map of CHOWDHARY et al. (2003) to increase the probability of detecting a QTL.
Fine mapping of the QTL identified will require the development of new markers for these regions, as no high-density equine marker maps are as yet available.
Summary
A set of 157 equine microsatellite markers was chosen and tested for its suitability for genome-wide QTL studies using 86 Hanoverian Warmblood horses. The markers were equally spaced over all 31 autosomes and the X chromosome, and the average distance was 19.1 cM. Multiplex sets were established in order to increase efficiency of genotyping. The marker set covered a total distance of 2081 cM, corresponding to 76.5% of the equine genome. Due to the small number of markers and the missing assignment of the markers to linkage groups, the density of markers was low on ECA07 and ECA28. The average observed heterozygosity of the whole marker set was 65.7%, the average expected heterozygosity was 64.6%, and the average polymorphism information content, 60.3%. The average PIC of the chromosomes ranged from 46.7% to 73.7%. The majority of the 157 markers (n = 123, 78.34%) was highly informative with a PIC > 50%. Therefore, this new microsatellite marker set provides a powerful tool for subsequent linkage studies in the horse.
Genome scan for QTL of OC in Hanoverian Warmblood horses using an optimised microsatellite marker set 37