Horsemap3Penedo

milies
and their genotypes were included in the analyses only to help identify pater-
nally transmitted alleles. Because of the paternal half-sib structure of map-
ping resources, except for NRF, only the recombination fractions for autoso-
mal markers obtained through male meioses were considered to construct the
comprehensive map.
The CRIMAP program version 2.4 (Green et al., 1990) was used for link-
age analysis. The program was modified to handle the increased number of
loci (n 1 99) and to adjust memory allocations. Maximum likelihood esti-
mates of recombination fraction (theta) were calculated using the TWO-
POINT option with a significant lod-score threshold 13 to determine the
appropriate linkage group association and chromosome assignment of 505
markers not previously included in the Phase II framework map (Guérin et
al., 2003). Multi-point analysis with the BUILD option was then used to
insert new loci within each chromosome, starting with a lod-score threshold
of 13. The best linear order was finally determined with a lod-score of 10.5
and checked with the FLIPS option. When the lowest negative log10_likeli-
hood was obtained for each of the 31 autosomes, the BUILD option was
rerun to check for additional insertions. This routine was repeated until the
best map was produced and no new marker appeared in the linear order. The
CHROMPIC option was used in a few instances to identify potential geno-
typing errors in the IHRFP data that could result in unlikely recombination
events, specifically, multiple recombinants within a small region. In six of
these cases, new genotype data were collected by retyping markers COR028,
COR058, COR062, COR100, LEX033 and UM004 at the Veterinary Genet-
ics Laboratory (code DAV). Map distances were calculated using the Kosam-
bi function restricted to the male-specific map.
Map representation
The linkage map was constructed based on the best final linear order and
distances. Cytogenetic information of physical assignments made by fluores-
cence in situ hybridization (FISH) was used to orient chromosomes. The
cytogenetic information reviewed, compiled and expanded by Chowdhary et
al. (2003) was used as the basic reference with additional information supple-
mented from other sources (http://locus.jouy.inra.fr/cgi-bin/lgbc/mapping/
common/intro2.pl?BASE=horse and http://www.thearkdb.org/).
Results and discussion
A total of 825 informative autosomal markers were ana-
lyzed, of which 766 (93%) were unambiguously assigned to one
of 31 linkage groups. Of these markers, 626 (82%) were linearly
ordered and 140 (18%) were assigned to a chromosomal region
but could not be inserted into the linear map with sufficient
statistical support. Fifty-nine markers (7%) failed to meet the
criteria for statistical evidence of linkage and remain unas-
signed. The total number of informative meioses (IM) was
182,467 with average number of IM of 248.48 B 135.16 for
ordered loci, 166.69 B 106.69 for assigned loci and 60.25 B
Cytogenet Genome Res 111:5–15 (2005) 7
49.88 for unassigned loci. Results from single factor ANOVA
for differences in average number of IM (data available in sup-
plement Table S3 at http://www.uky.edu/Horsemap/Consen-
susMap) between the three categories of marker mapping status
were highly significant (F = 75.27, df 2, 822, P = 9.6 × 10
–31
) as
were those between ordered and assigned (F = 44.76, df 1, 764,
P = 4.29 × 10
–11
). These analyses suggested that pooling the
mapping resources and the concomitant increase in number of
informative and co-informative meioses for common markers
had a positive effect in the ability to insert and order loci in the
map.
The linkage map spanning about 3,740 cM and oriented
according to the cytogenetic map is shown in Fig. 1. While the
number of markers in common between the two maps was suf-
ficient to help orient most linkage groups, many chromosomes
were poorly represented in the cytogenetic map (e.g. ECA5, 6,
7, 8, 9, 12, 13, 21, 24, 25, 28, 29, 30). An increase in the number
of physically mapped markers, as well as more dispersed FISH
assignments, would provide useful complimentary information
for accurate alignment of linkage groups and validation of the
order of loci obtained in this study.
The average distance between ordered markers is 6.3 B
5.8 cM (range 0–31 cM) with 56% of the intervals ^5 cM and
only 3% 620 cM. Details of coverage and map length by
chromosome are shown in Table 1. Relative to the previous
workshop map (Guérin et al., 2003), marker density increased
2.4-fold but the percentage of ordered and assigned loci
remained essentially the same at about 75 and 16%, respec-
tively. Seventeen chromosomes (ECA1, 2, 3, 4, 8, 9, 10, 12,
14, 15, 16, 18, 20, 22, 23, 30 and 31) contain one to three
regions with marker intervals between 18 and 31 cM with the
largest gaps located on ECA3 between HTG002 and TKY780
(31 cM) and ECA8 between TKY436 and UCDEQ046
(30 cM). Identification of informative markers to reduce inter-
vals and to improve coverage in the 17 chromosomes listed
above should be an important consideration for future horse
gene mapping research. This could be accomplished by tar-
geted mapping of new polymorphic microsatellites on the
IHRFP for which chromosome assignments are known from
twopoint linkage analyses on the NRF resource (Mickelson et
al., 2003; Swinburne et al., 2003) or RH panel (Wagner et al.,
2004) or by targeted development of markers using high reso-
lution RH maps to identify genes within the regions, select
BACs containing those genes and search for microsatellites
within those BACs.
With few exceptions, there was good agreement in linear
order of markers with previously published linkage maps. The
following differences are notable: VHL078 was ordered on
ECA2 (Guérin et al., 2003) but for the current map it was only
assigned to that chromosome as it could not be inserted in the
linear map with significant odds. The likely position of
VHL078 near ASB017 was suggested by results from TWO-
POINT analyses. HTG020 mapped to ECA7 based on signifi-
cant linkage to LEX038, LEX015 and TKY506, and not to
ECA5 as previously reported (Swinburne et al., 2000a).
NVHEQ067 mapped to ECA10 based on significant linkage to
TKY496, TKY503 and TKY722, and not to ECA22 as pre-
viously reported (Guérin et al., 2003). Furthermore, the current
Table 1. Marker density and average spacing between markers in the
horse linkage map for each chromosome
Chromosome Number of
ordered loci
Map span
(cM)
Average
spacing (cM)
Standard
deviation
Standard
error
ECA1 61 386.8 6.4 6.1 0.8
ECA2 30 174.2 6.0 5.7 1.0
ECA3 21 148.1 7.4 8.3 1.7
ECA4 35 142.8 4.2 4.4 0.7
ECA5 24 136.6 5.9 4.5 0.9
ECA6 20 123.2 6.5 5.0 1.1
ECA7 19 86.0 4.8 3.3 0.7
ECA8 27 154.6 5.9 7.4 1.4
ECA9 20 128.3 6.8 7.9 1.8
ECA10 30 163.4 5.6 5.8 1.1
ECA11 15 68.6 4.9 3.1 0.8
ECA12 10 70.2 7.8 5.8 1.8
ECA13 16 108.9 7.3 4.7 1.2
ECA14 23 168.4 7.7 7.5 1.6
ECA15 24 159.6 6.9 7.2 1.5
ECA16 29 151.5 5.4 5.5 1.0
ECA17 17 120.9 7.6 4.9 1.2
ECA18 15 154.2 11.0 7.3 1.9
ECA19 20 105.4 5.5 2.8 0.6
ECA20 17 142.2 8.9 5.4 1.3
ECA21 14 80.3 6.2 4.9 1.3
ECA22 16 99.6 6.6 6.5 1.6
ECA23 16 77.0 5.1 6.0 1.5
ECA24 22 87.5 4.2 3.6 0.8
ECA25 11 52.4 0.2 3.8 1.1
ECA26 13 80.4 6.7 5.5 1.5
ECA27 17 62.3 3.9 2.4 0.6
ECA28 13 74.9 6.2 5.1 1.4
ECA29 15 114.7 8.2 5.4 1.4
ECA30 8 64.0 9.1 8.4 3.0
ECA31 8 52.6 7.5 10.2 3.6
Total/Average 626 3739.6 6.3 5.8 0.2
assignment of NVHEQ067 agreed with the RH map assign-
ment (Chowdhary et al., 2003).
Discrepancies between FISH and linkage map locations
have been noted in published linkage maps for markers
AHT030, AHT044, ASB014, ASB038, SGCV008, SGCV032,
and TKY028. AHT030 was initially localized on ECA22 by
FISH (Swinburne et al., 2000b) but previous workshop linkage
data (Guérin et al., 2003) and this report (with increased num-
ber of meioses) placed this marker on ECA13. Incorrect FISH
locations of AHT044, ASB014 and ASB038 attributed to use of
chimeric clones as probes have been already been discussed
(Swinburne et al., 2000a). Published linkage maps and this
report supported location of these markers on ECA11, ECA8
and ECA27, respectively. Similarly, SGCV008 was physically
mapped to ECA19 (Godard et al., 1997) but linkage data for
this marker (Swinburne et al., 2000a; Guérin et al., 2003),
including this report, supported its location on ECA12. A new
FISH location of SGCV032 to ECA8 has been reported (Guérin
et al., 2003) that is now in agreement with linkage data. The
discrepancy for TKY028 has been noted (Guérin et al., 2003)
8 Cytogenet Genome Res 111:5–15 (2005)
Fig. 1. A male comprehensive linkage map of horse autosomes. Orientation of linkage groups is based on FISH assignments
shown on the cytogenetic map displayed next to G-banded schematic drawings of individual chromosomes on left (vertical lines
depict physical location). To optimize display, the anchor loci depicted are a subset of the markers in common with the linkage
map. Display of all common loci is available at http://www.vgl.ucdavis.edu/equine/caballus/. Idiogram nomenclature is taken from
the standardized karyotype (Bowling et al., 1997). Gray solid lines connect markers on both maps. Ordered markers are connected
to the bar diagram by a solid line. Markers not separated by recombination are shown with a branched marker position. Vertical
solid lines to the right of each linkage group indicate likely location of assigned loci. New markers added to the map are shown in
black font.
and linkage data available for this marker suppor