Theor Appl Genet (2007) 114:823–839
DOI 10.1007/s00122-006-0480-2
123
ORIGINAL PAPER
A 1,000-loci transcript map of the barley genome: new anchoring
points for integrative grass genomics
Nils Stein · Manoj Prasad · Uwe Scholz · Thomas Thiel · Hangning Zhang ·
Markus Wolf · Raja Kota · Rajeev K. Varshney · Dragan Perovic ·
Ivo Grosse · Andreas Graner
Received: 24 August 2006 / Accepted: 30 November 2006 / Published online: 12 January 2007
© Springer-Verlag 2007
Abstract An integrated barley transcript map (con-
sensus map) comprising 1,032 expressed sequence tag
(EST)-based markers (total 1,055 loci: 607 RFLP, 190
SSR, and 258 SNP), and 200 anchor markers from pre-
viously published data, has been generated by mapping
in three doubled haploid (DH) populations. Between
107 and 179 EST-based markers were allocated to the
seven individual barley linkage groups. The map covers
1118.3 cM with individual linkage groups ranging from
130 cM (chromosome 4H) to 199 cM (chromosome
3H), yielding an average marker interval distance of
0.9 cM. 475 EST-based markers showed a syntenic
organisation to known colinear linkage groups of the
rice genome, providing an extended insight into the
status of barley/rice genome colinearity as well as
ancient genome duplications predating the divergence
of rice and barley. The presented barley transcript map
is a valuable resource for targeted marker saturation
and identiWcation of candidate genes at agronomically
important loci. It provides new anchor points for
detailed studies in comparative grass genomics and will
support future attempts towards the integration of
genetic and physical mapping information.
Communicated by B. Keller.
Nils Stein and Manoj Prasad have contributed equally.
Electronic supplementary material The online version of this
article (doi:10.1007/s00122-006-0480-2) contains supplementary
material, which is available to authorized users.
N. Stein · M. Prasad · U. Scholz · T. Thiel · H. Zhang ·
M. Wolf · R. Kota · R. K. Varshney · D. Perovic ·
I. Grosse · A. Graner (&)
Leibniz Institute of Plant Genetics and Crop
Plant Research (IPK), Corrensstrasse 3,
06466 Gatersleben, Germany
e-mail: graner@ipk-gatersleben.de
Present Address:
M. Prasad
National Centre for Plant Genome Research (NCPGR),
JNU Campus, Aruna Asaf Ali Marg, Post Box No. 10531,
New Delhi 110067, India
Present Address:
H. Zhang
Agronomy Department, University of Florida,
IFAS, P.O. Box 11O300, Gainesville 32611-300, USA
Present Address:
M. Wolf
Trait Genetics, Am Schwabeplan 1b,
06466 Gatersleben, Germany
Present Address:
R. Kota
Plant Disease Resistance Group, CSIRO—Plant ,
Industry, Canberra, ACT 2601, Australia
Present Address:
R. K. Varshney
International Crops Research Institute for Semi
Arid Tropics (ICRISAT), Patancheru 502 324, AP, India
Present Address:
D. Perovic
Federal Centre for Breeding Research on Cultivated Plants,
Institute of Epidemiology and Resistance Resources,
Theodor-Roemer-Weg 4, 06449 Aschersleben, Germany
824 Theor Appl Genet (2007) 114:823–839
123
Introduction
Barley (Hordeum vulgare L.) is an important cereal
crop species ranking Wfth in crop production worldwide
after maize, wheat, rice, and soybean (area harvested,
FAO 2005, http://www.faostat.fao.org). The barley
genome (n = 7) comprising more than 5,000 Mb equals
approx. 12 times the size of the rice genome and con-
sists of about 80% of repetitive DNA (Flavell et al.
1974). Due to its importance as a staple crop and
because of its model character for other Triticeae
genomes including wheat, Triticum aestivum L. and
rye, Secale cereale L., comprehensive genetic and
genomic resources have been established for barley
over the past decades. These include a large number of
well-characterized genetic stocks and mutant collec-
tions (http://www.untamo.net/cgi-bin/ace/searches/basic)
(Caldwell et al. 2004; Lundqvist et al. 1996), various
genetic linkage maps (Varshney et al. 2004), large
insert bacterial artiWcial chromosome (BAC) libraries
(Isidore et al. 2005; Yu et al. 2000), and a large collec-
tion of expressed sequence tag (EST) presently com-
prising more than 4 £ 10
5
entries in dbEST (dbEST
summary July 21st, 2006, http://www.ncbi.nlm.nih.gov/
dbEST/dbEST_summary.html)
Expressed sequence tags obtained through cDNA
sequencing provide the link to gene information in
plant species, which are currently not suitable for
whole genome sequencing. In this regard, the available
EST collection, representing a large proportion of all
barley genes (Zhang et al. 2004), can be exploited for
barley in a similar way as has been shown for the con-
struction and anchoring of high-density genetic tran-
script and physical linkage maps in other plants such as
rice and maize (Chen et al. 2002; Davis et al. 1999;
Harushima et al. 1998; Kurata et al. 1997; Wu et al.
2002; Zhao et al. 2002). In bread wheat over 6,000
ESTs were allocated by deletion bin mapping to more
than 18,000 loci distributed across its allo-hexaploid
genome (http://wheat.pw.usda.gov/NSF/progress_map-
ping.html, Sorrells et al. 2003). This dataset provides a
starting point for the systematic analysis of gene/trait
associations, candidate gene identiWcation and compar-
ative genome analysis in grass species, keeping in mind
the limitation of low genetic and physical resolution
provided by the employed 159 cytogenetic mapping
bins (Qi et al. 2003).
Several approaches have been pursued for detecting
sequence polymorphisms in barley relying on hybridisa-
tion- (Restriction Fragment Length Polymorphisms,
RFLPs; e.g., Graner et al. 1991), or PCR-based molecu-
lar marker systems like RAPD (Randomly AmpliWed
Polymorphic DNA; e.g., Weyen et al. 1996), simple
sequence repeats (SSRs or microsatellites; e.g., Pillen
et al. 2000), ampliWed fragment length polymor-
phisms(AFLPs; e.g., Waugh et al. 1997) and single
nucleotide polymorphisms (SNPs; e.g., Kota et al.
2001b). Prior to the availability of PCR-based marker
techniques RFLPs have been most widely used because
of their simple development and their reproducibility.
Several detailed RFLP maps have been constructed in
barley comprising together more than 1,000 diVerent
markers (Kleinhofs and Graner 2001). RFLPs are usu-
ally inherited in a codominant way and bear the poten-
tial of parallel or subsequent multilocus mapping due to
cross hybridisation to independent gene family mem-
bers. Due to this feature, RFLP markers or their deriva-
tives facilitate an eYcient screening of BAC libraries
and provided the basis for the discovery of syntenous
relationships between plant genomes (Devos 2005; Hul-
bert et al. 1990). In case of cereals this facilitated access
to the fully sequenced genome of rice. However, their
detection is laborious and requires large amounts of
DNA especially in species with large genome size.
Hence, PCR-based SSR and SNP markers became the
preferred marker type in the past decade. SSRs are
stretches of DNA consisting of tandemly repeated short
units of 1–6 bp in length (Tautz 1989). Their polymor-
phic character arises due to variation in the number of
repeat units. They are multi-allelic and co-dominant in
nature and thus very informative (Powell et al. 1996).
EST databases can be mined for SSR containing ESTs
(for review see Varshney et al. 2005) allowing to obtain
markers at reduced cost for mapping of genes. On the
other hand SNPs are the most abundant form of genetic
variation and are less prone to mutations than SSRs
(Giordano et al. 1999). At genome-wide scale SNPs can
be expected at a frequency of 1/200–240 bp in barley
(Kota et al. 2001a; Rostoks et al. 2005). Computational
algorithms have been developed for querying EST data-
bases for the presence of SNPs (Kota et al. 2003), facili-
tating the systematic development of SNP markers, for
which innumerous assays have been developed (Rafal-
ski 2002; Wang et al. 1998).
High density genetic maps of gene-based markers
represent a powerful resource for enhanced genome
analysis. They are essential for linking genetic and phys-
ical mapping information and allow for a detailed com-
parative genome analysis across both closely related
and distantly related grass species. Moreover, gene-
based markers, also termed “functional markers,” can
be regarded as candidate genes in trait mapping experi-
ments. As a Wrst step towards a comprehensive tran-
script map of barley, more than 330 EST-derived SNP
markers were placed on a consensus map derived from
three mapping populations (Rostoks et al. 2005).
Theor Appl Genet (2007) 114:823–839 825
123
The aim of the present study was to further extend
the resource of mapped EST markers by developing a
high-density transcript map of the barley genome. To
maximize the potential of detecting polymorphisms,
diVerent marker technologies were employed, and
genetic mapping was performed in a genetically diverse
set of doubled haploid populations (Kota et al. 2001a).
Here, we report a genetic map of 1,055 loci detected by
1,032 EST-based markers. It provides a resource for
trait/gene association, candidate gene identiWcation,
marker saturation at independent target trait loci, and
represents a high density grid of entry points to the
genomes of rice and other grass species allowing a
reWned view onto grass genome colinearity and com-
parative genome organisation between rice and barley.
Materials and methods
Plant material
Three previously described doubled haploid (DH)
mapping populations were used in this study. Of these,
the population Igri £ Franka (I/F) (Graner et al. 1991)
was represented by 71 genotypes and the populations
Steptoe £ Morex (S/M) (Kleinhofs et al. 1993) and
Oregon Wolfe Dom £ Oregon Wolfe Rec (D/R)
(Costa et al. 2001), were represented by 94 genotypes
each. A comprehensive set of public marker data is
available (http://wheat.pw.usda.gov/ggpages/map_sum-
mary.html) for all three populations providing anchor
points for map integration and landmarks for map
comparisons within barley and to other grass species.
DNA markers
Expressed sequence tag sequences were obtained from
random sequencing of cDNA libraries developed from
a diverse set of tissues and developmental stages (Mic-
halek et al. 2002; Zhang et al. 2004) (CR-EST data-
base: http://pgrc.ipk-gatersleben.de/cr-est, Kuenne
et al. 2005) and served as source for EST-based marker
development (RFLPs, SSRs, SNPs). A tentative uni-
gene set was deWned by iterative clustering analysis
(project ID = g00¡g02 including between 13,000 and
111,000 ESTs; http://pgrc.ipk-gatersleben.de/cr-est)
using the software package StackPACK v2.1.1
(SANBI, South Africa). For genetic mapping, either a
singleton or a representative EST/cDNA-clone per
sequence contig was selected to avoid redundant map-
ping of genes. RFLP and SNP markers were randomly
selected from the EST collection, except of a subset
comprising about 60 SNP-markers, which were devel-
oped based on the identiWcation of SNPs present in the
public EST resource, which is derived from diVerent
genotypes (Kota et al. 2003). The development of SSR-
markers was based on pre-selecting ESTs containing
the corresponding repeat motifs (Thiel et al. 2003).
The developed markers were designated as GBR,
GBM and GBS (Gatersleben barley RFLP, microsatel-
lite and SNP) followed by a unique 4-digit numerical
identiWer. All mapped GB-markers were Wnally cross-
checked (BlastN, Altschul et al. 1990) for previously
unobserved redundancies against a unigene dataset
comprising over 370,000 publicly available EST
sequences (TIGR barley gene index release 9.0, 2004,
http://www.tigr.org/tigr-scripts/tgi/T_index.cgi?species =
barley).
Marker analysis
DNA extraction and Southern analysis were carried
out as described earlier (Graner et al. 1991) utilising a
set of six restriction enzymes (BamHI, Hin
dIII,
EcoRI, EcoRV, XbaI and DraI). Autoradiography
was performed by exposure of hybridised blots to
imaging plates (Fuji Photo Film, Japan) and subse-
quent signal detection on a phosphoimager (FLA-
3000, Fuji, Japan). cDNA inserts were ampliWed by
utilising standard sequencing primers, puriWed
(Qiaex; Qiagen, Hilden, Germany), radioactively
labelled (according to manufacturers instructions:
Megaprime labelling system; Amersham Biosystems,
Freiburg, Germany) and utilised as RFLP probes
according to Graner et al. (1991). The development
and analysis of EST-based SNP and SSR markers fol-
lowed previously published protocols (Kota et al.
2001b; Thiel et al. 2003). Detailed information (NCBI
Genbank accession number of underlying EST, chro-
mosome location, consensus map position, primer
sequences in case of PCR-markers) is provided as
Electronic Supplementary Material (ESM) Table 1.
Primer info for markers GBM1001-1076 is available
based on an MTA upon request to the corresponding
author.
Linkage analysis and map construction
Genotyping information was recorded for each marker
by entering segregation data into population Wles utilis-
ing the software
MAPMANAGER QTX v0.30 (Manly et al.
2001). These Wles included previously published
marker data (see below) thus allowing to Wt new
marker data into the seven barley linkage groups using
the command “Distribute” (LOD 3.0 for I/F, and LOD
4.0 for S/M and D/R).
JOINMAP V3.0 (Kyazma, The
826 Theor Appl Genet (2007) 114:823–839
123
Netherlands) was used for grouping of markers (LOD
score = 4.0) and subsequent determination of marker
order (minimum LOD score = 1.0, recombination
threshold 0.4, ripple value = 1, jump threshold = 5).
The Kosambi mapping function (Kosambi 1944) was
applied for converting recombination units into genetic
distances. Graphical genotypes of the resulting individ-
ual chromosome maps were visually inspected for con-
sistency. In order to avoid a contradictory placement of
loci (i.e., new double crossing-over introduced due to
false marker order) that occurred occasionally, individ-
ual maps were recalculated by setting individual loci at
‘Wxed order’. Map integration (consensus map) was
performed with
JOINMAP V3.0 under the conditions/set-
tings as described above applying the Kosambi map-
ping function (Kosambi 1944) for converting
recombination units into map distances. The marker
order of the consensus chromosome maps was com-
pared to the original order in the individual population
maps. In six cases blocks of markers spanning at maxi-
mum 3 cM displayed an inverted order compared to
the map of the individual population thus violating the
original graphical genotype. Here the consensus map
was hand-curated to conform to the marker order sup-
ported by experimental evidence.
Mapping data of 200 previously published markers
(ESM Table 2) was utilised as a framework for building
the consensus map. These markers originated from
various laboratories and included apart from morpho-
logical and isozyme loci a majority of DNA-based
molecular markers originating from cDNA (BCD and
CDO, Heun et al. 1991; cMWG, Graner et al. 1991;
ABC, Kleinhofs et al. 1993; Bmac, Ramsay et al. 2000)
or genomic clones (MWG, Graner et al. 1991; ABG,
Kleinhofs et al. 1993; WG, Heun et al. 1991; and Ksu,
Gill et al. 1991), or miscellaneous clones (ABA, Klein-
hofs et al. 1993). The approximate position of the cen-
tromeres was determined according to Kuenzel et al.
(2000). Final chromosome maps were drawn with the
graphical package MAPCHART (Voorrips 2002).
All mapping data (individual maps, consensus map,
comparative map) can be visualized through internet
(http://pgrc.ipk-gatersleben.de/transcript_map) by uti-
lizing the visualisation tool MoMaVis.
IdentiWcation of orthologous genes in the rice genome
Expressed sequence tags of the 1,032 experimentally
mapped barley cDNAs were aligned with the publicly
available rice genome sequence (TIGR, http://www.
tigr.org/tdb/e2k1/osa1/, version 3, February 18, 2005)
by BlastN (E · 1E-10). The genetic map positions of
the barley genes were plotted against the physical
coordinates of their best homologs (putative orthologs)
from rice (Fig. 4) in order to determine the subset of
syntenic genes between barley and rice.
Inferring barley duplications
Barley chromosomes 2H and 6H carry colinear regions
to rice chromosomes Os04 and Os02, which were
involved in an ancient whole genome duplication in
rice (Yu et al. 2005) predating the species divergence
of barley and rice. Syntenic regions are based on Wnd-
ing the putative ortholog for a mapped barley EST,
which is deWned as the rice gene with the BlastN align-
ment with the lowest E value (E · 1E-10). In order to
Wnd putative paralogs between barley chromosomes
2H and 6H and rice chromosomes Os02 and Os04,
respectively, second-best rice homologs were addition-
ally extracted. To examine whether second-best BlastN
hits were signiWcantly accumulated in these syntenic
regions, a one-sided Fisher’s exact test was used to test
the null hypothesis of no association between the vari-
ables “located on rice chromosome x” and “located on
barley chromosome y”. The distribution of best and
second-best rice homologs was studied with the same
test on the null hypothesis assuming no correlation
between ESTs from barley chromosome 2H and chro-
mosome 6H ESTs and the distribution of their corre-
sponding best and second-best rice homologs across
the rice chromosomes Os02 and Os04. In both cases
the null hypothesis was rejected if P · 0.05.
Results
Analysis of RFLP markers
An overview of the overall number and characteristics
of all newly derived EST-based RFLP, SSR and -SNP
markers is provided in Table 1 and 2. If compared
across all three populations, SNP-markers represented
the most polymorphic class of markers: 57% detected a
polymorphism as compared to 51 and 38% for RFLPs
and SSRs, respectively (Table 1).
For the development of RFLP-markers cDNA
clones were selected based on their corresponding EST
sequence and tentative unigene information. Overall,
1,539 out of 1,578 clones (97.5%) showed clear and
useful hybridisation signals among the parents of the
employed mapping populations (Table 1). 539 RFLP
markers out of 782 polymorphic probes were Wnally
mapped detecting 555 loci. Together with previously
characterized cDNA markers (Graner et al. 1991) a
total of 584 EST-based RFLP-markers detecting 607
Theor Appl Genet (2007) 114:823–839 827
123
loci were included for map construction. Altogether,
between 53 and 107 RFLP loci could be assigned to
each of the seven barley chromosomes. Out of the 607
loci, 168, 172, and 295 were mapped in the populations
I/F, S/M, and D/R, respectively (Table 2) with 13 mark-
ers detecting either two and 5 markers detecting three
polymorphic loci, respectively (ESM Table 3).
Analysis of SSR and SNP markers
A set of 190 EST-SSR markers (including 185 previ-
ously published; Varshney et al. 2006) as well as 258
SNP markers (including 221 to be published else-
where) were analysed as described before (Kota et al.
2001b; Thiel et al. 2003; Varshney et al. 2006) and the
results were integrated together with the RFLP data
into a combined barley transcript map (Fig. 1).
Construction of a transcript map
Individual genetic maps were calculated for each of the
three DH populations (I/F, S/M and D/R) preceding
the integrated map construction. 585 loci were mapped
in D/R, 311 in S/M and 209 in I/F (Table 2). Further-
more, segregation data of 200 published markers
(Costa et al. 2001; Graner et al. 1991; Kleinhofs et al.
1993; http://wheat.pw.usda.gov/ggpages/map_summary.
html) was included (ESM Table 2) to provide a frame-
work for the construction of the consensus map and to
serve as points of reference to previously published
maps. The observed order of anchor markers in the
computed individual maps was in accordance with
previously published maps.
Subsequently, a consensus transcript map was calcu-
lated (Fig. 1) comprising 1,032 EST-based marker
Table 1 Aggregated information on polymorphism for diVerent types of EST-based markers
a
Excluding cMWG markers
b
R. Kota et al., unpublished data
c
Varshney et al. (2006)
d
17, 6, and 16 RFLP-, SNP-, SSR-markers were mapped in two populations, respectively. One multi-locus RFLP-marker was mapped
at one locus in two and at a second locus in all three populations
Assay employed RFLP
a
SNP
b
SSR
c
Screened ESTs 1,578 710 759
Potential candidates 1,539 436 532
Polymorphism detected 782 (51%) 264 (57%) 201 (38%)
Polymorphic in I/F 246 (16%) 74 (17%) 58 (11%)
Polymorphic in S/M 452 (29%) 158 (36%) 107 (20%)
Polymorphic in D/R 518 (34%) 193 (45%) 155 (30%) Total
Mapped in I/F 114 18 23 156
Mapped in S/M 162 92 47 302
Mapped in D/R 282 154 136 572
Total non-redundant 539
d
258
d
190
d
990
d
Table 2 Summary of EST-based marker loci for the individual maps and the integrated consensus map
a
Number of redundant EST mapped either by RFLP (including cMWG), SSR and SNP
b
18 RFLP detected 23 secondary or tertiary loci (607 loci—23) = 584 EST-RFLP probes used
c
Represented by 1,032 non-redundant markers
Chromosome Population/map
I/F S/M D/R Integrated map
a
No. of loci Map
length
No. of loci Map
length
No. of loci Map
length
No. of loci Map
length
RFLP SNP SSR Total (cM) RFLP SNP SSR Total (cM) RFLP SNP SSR Total (cM) RFLP SNP SSR Total (cM)
1H 22 2 5 29 132.5 28 9 5 42 122.3 38 17 20 75 133.0 84 27 27 138 134.3
2H 24 1 1 26 133.7 35 18 8 61 146.7 56 21 25 102 174.5 107 39 33 179 165.1
3H 27 7 3 37 137.6 32 14 11 57 157.6 48 24 24 96 210.3 101 43 35 179 199.3
4H 9 0 1 10 137.8 9 10 7 26 129.7 36 20 20 76 130.5 53 28 26 107 129.8
5H 23 2 2 27 187.1 29 12 4 45 153.8 50 35 18 103 222.3 98 49 23 170 197.2
6H 22 3 7 32 129.2 17 12 5 34 108.9 25 17 17 59 143.0 60 32 25 117 149.7
7H 41 3 4 48 168.3 22 17 7 46 133.5 42 20 12 74 150.4 104 40 21 165 142.9
Total 169 18 23 210 1,026.2 173 92 47 312 952.5 296 154 136 586 1,164.0 607
b
258 190 1,055
c
1,118.3
