Genetic Resources and Crop Evolution
Analysis of genetic diversity in Tunisian durum wheat cultivars and related
wild species by SSR and AFLP markers
Maher Medini1, Sonia Hamza1,*, Ahmed Rebai2and Michael Baum3
1Laboratory of Genetics and Plant Breeding, Institut National Agronomique de Tunisie, Avenue Charles
Nicolle, Tunis 1082, Tunisia;2Laboratory of Plant Protection and Transformation, Centre de
Biotechnologie de Sfax, BP ‘K’, Sfax 3038, Tunisia;3ICARDA, Germplasm Program, P.O. Box 5466,
Aleppo, Syria; *Author for correspondence (e-mail: firstname.lastname@example.org)
Received 29 October 2002; accepted in revised form 30 May 2003
Key words: AFLP, Durum wheat, Genetic diversity, Microsatellite, SSR, Triticum durum
Thirty-four durum wheat cultivars representing the Tunisian durum (Triticum durum Desf.) wheat collection
and seven wild species of wheat relatives (Triticum turgidum L., T. dicoccon Schrank., T. dicoccoides (Ko ¨rn)
Schweinf., T. araraticum Jakubz., T. monococcum L., Aegilops geniculata Roth, and Aegilops ventricosa
Tausch) were analysed with amplified fragment length polymorphism (AFLP) and microsatellite (SSR)
markers. Both marker systems used were able to differentiate durum wheat cultivars from the wild relatives
and to specifically fingerprint each of the genotypes studied. However, the two marker systems differed in the
amount of detected polymorphisms. The 15 SSR markers were highly polymorphic across all the genotypes.
The total number of amplified fragments was 156 and the number of alleles per locus ranged from 3 to 24
with an average of 10.4. Two SSR markers alone, Xwms47 and Xwms268, were sufficient to distinguish all 34
durum wheat genotypes. The five AFLP primer pair combinations analysed yielded a total of 293 bands, of
which 31% were polymorphic. The highest polymorphic information content (PIC) value was observed for
SSRs (0.68) while the highest marker index (MI) value was for AFLPs (7.16) reflecting the hypervariability
of the first and the distinctive nature of the second system. For durum wheat cultivars, the genetic similarity
values varied between 31.3 and 81% for AFLPs (with an average of 54.2%), and between 3.6 and 72.7% for
SSRs (with an average of 19.9%). The rank correlation between the two marker systems was moderate, with
r ¼ 0.57, but highly significant. Based on SSR markers, highest genetic similarity (GS) values were observed
within the modern cultivars (37.3%), while the old cultivars showed a low level of GS (19.9%). Moreover, the
modern cultivars showed low PIC and MI values. UPGMA Cluster analysis based on the combined AFLP
and SSR data separated the wild wheat species from the durum wheat cultivars. The modern cultivars were
separated from the old cultivars and form a distinct group.
Wheat is the most important cereal crop in Tunisia
where it covers 60% of the cereal area. Prior to
1969, wheat breeding was based on the exploitation
of the variability in the local durum wheat popula-
cultivars were developed by mass and pedigree
selectionfromsegregantpopulations(Degha€ ? ?setal.
1998). Population increase and food consumption
led to changes in the breeding strategy in order to
develop varieties more productive and suitable
to intensive agriculture. Breeding programs using
exotic germplasm resulted in the development of a
number of modern cultivars, which replaced the
former heterogenous populations. By shifting to
# Springer 2005
52: 21–31, 2005.
more productive cultivars the national yield
increased from 500 kg/ha (1930–1970) to 1280 kg/
At present, Tunisian durum wheat germplasm
does not exceed 45 varieties and it consists of: (i)
old varieties selected from local populations (1890–
between old varieties (1930–1970); and (iii) modern
cultivars bred at the international breeding centers
and further selected in Tunisia (1970–today).
Tunisian durum wheat germplasm has been well
characterized for morphological, physiological and
agronomic traits and it has been shown that local
germplasm is diverse and it is a source of genes for
resistance/tolerance to biotic and abiotic stresses as
well as technological qualities (Bœuf 1932; Rezgui
et al. 1998; Karmous et al. 2002). However, to date,
no information was available on genetic variation
of Tunisian durum wheat germplasm at the mole-
cular level. Genetic diversity estimation based on
DNA markers is more informative when the mar-
kers used are distributed on the whole genome.
PCR based markers such as RAPD, AFLP
and SSR, have been shown to be powerful tools for
durum wheat cultivars (Dograr et al. 2000; Eujayl
et al. 2002; Soleimani et al. 2002; Pujar et al. 1999;
Incirli and Akkaya 2001). The objectives of the
present study were to evaluate the genetic diversity
of durum wheat in a representative set of the
Tunisian national collection and to differentiate
durum wheat collections from wild relatives by
utilising AFLP and SSR markers.
Materials and methods
Plant material and DNA isolation
Forty tetraploid and one diploid wheat genotypes
were used in this study (Table 1). They consisted of
34 durum wheat cultivars of the Tunisian national
collection, comprising 25 old varieties selected
from local and foreign populations, two intermedi-
ate varieties from a cross between old cultivars, and
the whole collection of seven modern cultivars
released between 1980 and 1996. Seven accessions
of wild related species that are commonly used for
hybridisation and development of intermediate and
new varieties were provided by the International
Center for Agricultural Research in the Dry Areas
(ICARDA). Additionally, two Aegilops species
found in Tunisia (Ae. geniculata and Ae. ventricosa
accessions collected from the North (Ariana) and
the Center (Kef) of Tunisia, respectively) were
included in the analysis, as they are the most wide-
spread wild wheat relatives in Tunisia, possibly
harbouring useful disease resistance genes for
Total genomic DNA was isolated from fresh leaf
material (ten plants per genotype) at the adult stage
Maroof et al. (1984). DNA concentration was esti-
mated using a spectrophotometeras well as by visual
comparison with known concentrations of phage
AFLP analysis was performed as described by
Barrett and Kidwell (1998). One microgram of
genomic DNA was digested with restriction
enzymes PstI and MseI and double stranded
adaptators ligated to the fragment ends. This was
tive primers. Selective amplification was performed
on the pre-amplified fragment mixture using a total
of five PstI:MseI primer combinations (Table 2).
adjusted to 10 ng/?L. Fifteen primer pairs described
by Ro ¨der et al. (1998) were used in the present study
(Table 3). PCR amplifications were carried out in a
Perkin Elmer thermocycler model (PE9700) in a
25 ?L reaction mixture. Each reaction contained
1? Taq polymerase buffer, 200 ?M each dNTP,
0.5 ?M of each of the two primers, 1 U Taq poly-
merase (Promega), and 80 ng of genomic DNA as
94?C for 2 min, 40 cycles of (94?C 30 s, 55 or 60?C
30 s, 72?C 45 s), 72?C for 10 min.
For both assays (AFLP and SSR), PCR
products were mixed with loading buffer, 5–10 ?L
of mixture were denatured and loaded into wells in
0.4 mm thick 6% denaturing polyacrylamide
gels (35?50 cm) resolved at constant power
(75 W) in 1?TBE running buffer for 2.15 h.
Table 1. Genotypes used in AFLP and SSR analyses, their genome, their pedigree (when available), their breeding institute or region of
origin and the year of introduction or release (in case of inbred cultivars).
Species Genome VarietiesPedigree x
¼ BR‘S’ ¼ BEL116E/2*TC60//ZB/WLS
¼ D58-25 ¼ Kyperounda?Mahmoudi 981
D56-3A ¼ Zenati Bouteille//Mahmoudi 152/Mrari
Wild wheat species
*Inbred lines; x Jori ¼ JO‘S’; ZB ¼ Zenati Bouteille; ¥ CIMMYT: The International Maize and Wheat Improvement Center; ICARDA:
International Centre for Agricultural Research in the Dry Areas; INRAT: Institut National de Recherche Agronomique de Tunisie.
Bands were visualised by the silver-staining method
as described by Pillen et al. (2000). Dried gels were
visually scored and scanned for records.
To measure the informativeness of the markers, the
average polymorphic information content (PIC)
was calculated for each marker system according
to the formula given by Powell et al. (1996):
PIC ¼ 1 ?
where fiis the frequency of the ith allele in the set
of 41 genotypes. This value provides an estimate
of the discrimination power of a molecular locus
by taking into account not only the number of
alleles per locus but also their relative frequencies
in the studied population (L€ u ubberstedt et al.
The marker index (MI) is determined in order to
balance the level of polymorphism detected (PIC)
and the extent to which an assay can identify multi-
ple polymorphisms. It was calculated according to
MI ¼ PIC ? ? ? ?
where ? is the proportion of polymorphic bands
and ? is the number of loci per assay unit (Powell
et al. 1996).
Estimates of genetic similarity
Data obtained with the two assays were scored in a
binary form as presence (1) or absence (0) of bands
for each individual, and entered into a data matrix.
Genetic similarity (GS) between two genotypes i
and j was calculated for each marker system and
Table 2. Oligonucleotide sequences used in the AFLP analysis.
PstI Forward adapter
50CTC GTA GAC TGC GTA CAT GCA 30
50TGT ACG CAG TCT AC 30
50GAC GAT GAG TCC TGA G 30
50TAC TCA GGA CTC AT 30
PstI-A50GAC TGC GTA CAT GCA GA 30
MseI-C50GAT GAG TCC TGA GTA A C 30
PstI-ACG50GAC TGC GTA CAT GCA G ACG 30
PstI-ACA50GAC TGC GTA CAT GCA G ACA 30
PstI-AGG50GAC TGC GTA CAT GCA G AGG 30
PstI-AAG50GAC TGC GTA CAT GCA G AAG 30
MseI-CGC50GAT GAG TCC TGA GTA A CGC 30
MseI-CAA50GAT GAG TCC TGA GTA A CAA 30
MseI-CAC50GAT GAG TCC TGA GTA A CAC 30
Table 3. List of SSR markers used, their map position, primer sequence according to Ro ¨der et al. (1998), and their number of alleles,
PIC and MI values generated in the present germplasm.
SSR locus Chromosome(s) MotifNumber of alleles PICMI
across marker systems according to the formula
given by Nei and Li (1979):
GSij¼ 2Nijl ðNiþ NjÞ
where Nij is the number of bands in common
between cultivars i and j, and Ni and Nj are
the total number of bands in genotype i and j,
respectively, with regard to all assay units. Thus
GSijreflects the proportion of bands in common
between two genotypes and may range from 0
(no common bands) to 1 (identical profiles of two
Coefficient of co-ancestry
Coefficients of co-ancestry (f) were calculated for
pair of the seven modern cultivars from the
pedigree records described in Table 1. f were com-
puted for all pairs of lines as fellows (Kempthorne
where f between the A and B lines is equal to the
average of fs between A and the parental lines of B
(C and D).
The calculations followed Cox et al. (1985)
assumptions: (1) ancestors are unrelated; (2) all
cultivars, ancestors and parental lines are homozy-
gous; and (3) a cultivar derived from a cross
obtains one-half of its genes from each parent.
The similarity matrices obtained with the two sets
of data – SSR and AFLP – were converted to
distance measures (d) using a formula d ¼ 1?GS
and used to generate UPGMA (unweighted pair
group method with arithmetic averages, Sneath
and Sokal 1973) dendrograms, visualised with
Treecon (Van de Peer and De Wachter 1994).
Estimation of correlation coefficient
Correlation between similarity matrices derived
from SSR and AFLP was tested by Mantel’s test
(Mantel 1967) with the NTSYS-PC (Version 2.1)
program (Rohlf 2000). We also computed linear
(denoted r) and Spearman rank (denoted rs) corre-
lation coefficients between GS values of AFLP
and SSR. Significance P-values for these correla-
tions were calculated based on their finite sample
matrices GS was tested with the Mantel Z statistic
(Mantel 1967). Significance of Z was determined
by comparing the observed Z values with a critical
Z value obtained by calculating Z for one matrix
with 1000 permuted variants of the second matrix.
All computations were performed with appropriate
procedures of the NTSYS-PC version 2.1 software
Five AFLP primer combinations yielded a total of
293 amplification products, 91 of them were poly-
morphic across all the 41 genotypes with an aver-
age of 18.2 polymorphic bands (Table 4). Each
primer combination was able to distinguish the 41
genotypes. When considering either of the durum
the average number of polymorphic bands was 17.2
and 17, respectively. However, this value decreased
dramatically within the modern cultivars (12.9).
A total of 156 fragments were obtained from the 15
SSR primers and all the bands were polymorphic
across all the genotypes screened. The number of
alleles per locus ranged from 3 to 24, with an
average of 10.4 alleles per locus (Table 3). When
considering durum wheat cultivars, 14 of the 15
used SSR primers yielded 112 polymorphic bands
with an average of eight alleles per SSR locus. This
average was similar within old cultivars (7.43), but
it fell to 3.17 when considering just modern culti-
vars (Table 4).
The maximum number of alleles was detected
at Xwms268 (Table 3). This SSR marker, which is
mapped to a single locus on chromosome 1B
(Ro ¨der et al. 1998), showed two PCR products on
17 genotypes. However, the Xwms497, which is
mapped to two loci (1A and 2A), showed a single
monomorphic band for all durum wheat geno-
types. Two microsatellite primer pairs, WMS268
and WMS47, were sufficient to differentiate all the
durum wheat genotypes.
SSR markers that mapped either to chromo-
QJ;somes of the homologous groups A or B were
also amplified in wild species carrying a different
genome. Among eight markers mapped on the
(MMUU)). B-genome SSR markers amplified
four and three bands and A-genome SSR markers
amplified two bands
Ae. geniculata, respectively. However, other A-genome
SSR markers did not amplify any PCR product in
species carrying the A genome. Among the eight
markers mapped in this genome, only three
(Xwms60, Xwms445, and Xwms674) were amplified
in diploid species carrying the A genome (T. mono-
coccum) and six (Xwms60, Xwms304, Xwms445,
Xwms570, Xwms637, and Xwms674) were amplified
in the tetraploid species carrying A+G genomes
Across all the analysed genotypes, the average
polymorphic information content values were dif-
ferent between the two marker systems (Table 4).
The SSRs have shown higher PIC values than the
AFLPs (0.72 and 0.39, respectively). However, the
MI was definitely higher for AFLPs than for SSRs
(7.16 and 0.81, respectively), which reflects the
ability of the AFLP technique to discriminate by
few primer combinations all the genotypes tested.
For the SSR assays, the PIC values were different
when considering separately old and modern
cultivars (respectively, 0.68 and 0.44), whereas the
AFLP assays generated close PIC values (0.35,
and 0.38, respectively) in the same materials.
Nevertheless, differences were detected between
the MI values (5.9 and 4.9, respectively) for AFLP.
Similarity matrices constructed on shared allele
bases revealed that the average genetic similarity
between genotypes was lower using SSRs (15% for
all genotypes and 19.9% for durum wheat cultivars;
Table 5) than AFLPs (39.4% for all genotypes
and 54.2% for durum wheat cultivars). With AFLP
markers, GS average values were quite similar in
old and modern cultivars (50.5 and 57.7%, respec-
tively), whereas with SSR markers, modern culti-
vars GS was higher than that for old cultivars (37.3
and 19.9%, respectively).
Coefficients of co-ancestry (f) were calculated for
21 pairs of modern cultivars (data not shown). fs
values ranged from 0.000 to 0.500 with a mean of
In order to illustrate the relatedness between the 41
wheat genotypes, a combined UPGMA cluster
Table 4. Analyses of banding patterns generated by SSR and AFLP assays.
assay unit PIC MI
SSRAll genotypes 4115 (primer
15610.41 10.41.13 0.72 0.81
341128.01 8.01.07 0.68 0.72
3429358.6 0.29 17.2 58.60.38 6.5
based on the GS matrix calculated by the combina-
tion of 247 polymorhic bands obtained with SSR
and AFLP data was developed (Figure 1). The
cluster analysis clearly separated the wild wheat
species from the durum wheat cultivars. The
modern cultivars were separated from the old culti-
vars and form a distinct group. Indeed, the modern
cultivars are more productive, insensitive to day-
length, earlier than old cultivars and with exception
of ‘Amel’, they are related by pedigree (Table 1).
Table 5. Maximum, minimum and mean genetic similarity estimates (expressed in percentage) calculated from AFLPs and SSRs for all
the genotypes (41), durum wheat cultivars (34), old cultivars (25) and modern cultivars (7). Coefficient of Co-ancestry was calculated
only for modern cultivars.
Number of genotypes
MaxMin Mean MaxMin Mean MaxMinMean
37.3 0.5000.000 0.053
Figure 1. Dendrogram of 34 durum wheat cultivars and seven related species revealed by UPGMA cluster analysis based on genetic
similarity (GS) estimates calculated from the combined AFLP and SSR data.
The old cultivars were clustered into two majors
groups. However ‘Zenati Bouteille’ and ‘Derbessi’
were not included. Despite the fact that the old
cultivar ‘Zenati Bouteille’ has been used to breed
the cultivars ‘Amel’ and ‘Badri’ (see pedigree), this
cultivar is well separated from its progeny.
Comparison between SSR and AFLP
To compare the results obtained with the two tech-
niques, the correlation was tested using linear (r)
correlation and Spearman rank (rs) correlation of
the pair wise GS values. Correlations between both
GS matrices were highly significant (r ¼ 0.57, P <
0.0001; rs ¼ 0.63, P < 0.0001). The normalised
Mantel statistic Z test was also significant between
GS matrices (Z ¼ 0.678, P < 0.002).
The two DNA marker techniques used in the pre-
sent study were able to clearly discriminate between
the 41 cultivated and wild wheat genotypes studied.
Using AFLP, only five primer pair combinations
were sufficient to generate 91 polymorphic mar-
kers, revealing differences between related modern
cultivars. A total of 156 fragments were obtained
from the 15 SSR primers amplified and all bands
were polymorphic across all the genotypes used.
The two microsatellite primer pairs, WMS268 and
WMS47 alone were sufficient to differentiate all the
durum wheat genotypes.
The present study shows that the polymorphic
information content (PIC) for SSRs is greater than
for AFLP. This is related to the variation of the
average number of polymorphic bands per assay
unit. This average ranged from a maximum of
100% for SSR to only 31% for AFLP. The high
amount of variability has also contributed to the
lower estimates of GS based on SSR markers
(mean 19.9%) as compared to AFLP markers
(mean 54.2%). Our results are in agreement with
other studies comparing the level of polymorphism
detected with AFLP and SSR markers in wheat
(Bohn et al. 1999; Donini et al. 2000), barley
(Russell et al. 1997), maize (L€ u ubberstedt et al.
2000), hops (Jak? s se et al. 2001), and in soybean
(Powell et al. 1996). However, the calculation of
the marker index (introduced by Powell et al. 1996)
highlights the distinctive nature of the AFLP assay:
the high marker index or diversity index is a reflec-
tion of the efficiency of AFLPs to simultaneously
analyse a large number of bands rather than the
level of polymorphism detected.
Relative to other diversity studies on wheat,
more alleles (an average of 15 alleles) were detected
in our study at loci containing (GA) repeats. Ro ¨der
et al. (1995) also reported (GT) repeats to be more
polymorphic than other repeats in wheat, whereas,
Akagi et al. (1997) found that the highly poly-
morphic microsatellites of rice consist of (AT)
repeats. The average number of eight alleles per
locus for durum wheat cultivars recorded in the
present study was relatively higher than those ear-
lier reported on durum wheat cultivars (average of
6.5 in Dograr et al. 2000 and 4.5 in Eujayl et al.
2002). This difference might relate to the utilisation
of old cultivars with a mean value of 7.43 alleles per
locus versus 3.17 among modern cultivars as well as
the selection of the specific primer.
Specifically, we have compared the genetic varia-
bility between old durum wheat cultivars and mod-
ern cultivars. Genetic similarity estimates based on
SSR allele polymorphism differed between the two
sets of cultivars in contrast to GS estimates based
on AFLP data. Based on SSR data, modern culti-
vars were genetically more similar to one another
than old ones. This may suggest that a high pro-
portion of SSR markers used in our study might be
associated with chromosomal regions selected dur-
ing breeding programs that have more fixed allelic
state. This could be chromosomal regions harbor-
ing a number of agronomic important loci (such as
those determining vernalisation requirement, semi-
dwarf habit or yield) that are more uniform in the
modern varieties. Microsatellite frequency was
shown to be higher in transcribed regions, espe-
cially in the untranslated portions, than in other
genomic DNA fractions (Morgante et al. 2002) and
have a significant association with the low-copy
fraction of plant genomes. Therefore, SSR markers
were more under the influence of selection, adapta-
tion and use than neutral AFLP markers not
affected by selection (Stachel et al. 2000; Saghai-
Maroof et al. 1994). The more widespread applica-
tion of SSR markers, especially EST-derived SSR
markers to the evaluation of germplasm collection
will provide an opportunity to examine functional
diversity in phenotypically well characterized
accessions (Eujayl et al. 2002). Old varieties will be
the appropriate material for genetic diversity stu-
dies in relation to the environment adaptation since
their adaptation is secular (Bœuf 1932).
The SSR markers used in our study were not
specific to a single genome. Several markers ampli-
fying loci of the A or B genome also amplified loci
of Ae. ventricosa and Ae. geniculata, carrying the
UM and DN genomes respectively. Microsatellites
of bread wheat were transferable to homoeologous
species (Guadagnuolo et al. 2001), however,
this transferability was far weaker than that
observed from wheat to the corresponding ances-
tral diploid species genome (Sourdille et al. 2001).
The recent findings of Morgante et al. (2002) show-
ing a significant association of SSR with the low-
copy fraction of plant genomes may also suggest a
considerable amount of conservation of micro-
satellite and microsatellite flanking sequences
between species which might be exploited for com-
parative genomic analysis.
Genetic relatedness between genotypes was
tested using UPGMA cluster analysis. A dendro-
gram based on the combination of AFLP and SSR
data with a total of 247 polymorphic markers was
developed in order to describe the relation-
ship between the varieties. Interestingly, we have
noticed that all the modern cultivars were sepa-
rated from old cultivars and were clustered in one
group. Despite the fact that most of the modern
cultivars are related by pedigree, they share several
agro-morphological parameters that are different
from those observed in old cultivars (plant height,
yield, earliness and insensitivity to day-length). The
insensitivity to the day-length is a consequence of
sowing twice a year in the south and north of
Mexico by the CIMMYT wheat program. This
phenomenon would largely contribute to the dis-
tinction of the modern from the old varieties
(Gharbi pers. comm). Some groups obtained in
this study were consistent with the classification
based on morphological spike
(Bœuf 1932). Indeed, the subgroup composed of
the varieties ‘Bidi, ‘Roussia’, ‘Hamira’, Jenah
Khottifa’, and ‘Souri’ share the similar colour
trait of the glumes and awn (russet), and barbs
(russet or black). The relatedness between ‘Agili’,
‘Sbei’, ‘Biadha RP1’, ‘Adjini’, ‘Mekki’ and ‘Biskri’
is supported by similar spike characteristics (white
glumes and black or white barbs).
In our study associations were found between
GS-AFLP and GS-SSR with a pool of related and
unrelated varieties. According to Bohn et al. (1999)
correlation between different marker estimates is
observed only for related cultivars. The combi-
nation of unrelated parent correlations between
GS estimates determined with different marker sys-
tems can be expected only if linkage disequilibrium
exists between the different marker loci in the gen-
otypes examined (Bohn et al. 1999). In our study,
the chance for close linkage between AFLP and
SSR marker loci was low since only 15 SSR and
92 polymorphic AFLP markers were used with low
probability to be tightly linked. Therefore, one
reason that may account for the association
observed between GS measures is that linkage dis-
equilibrium exists among the cultivars forming
three different groups (wild, old and modern
wheat) as it was demonstrated by the dendrogram
(Bohn et al. 1999).
and GS estimates was not significant despite the
fact that modern cultivars were related by pedigree
(data not shown). In our case, this lack of correla-
tion can be attributed to the low number of culti-
vars with known pedigree (7 cultivars). However,
low correlations between f and GS based on mole-
cular markers were also reported in other studies
(Plaschke et al. 1995; Autrique et al. 1996; Bohn
et al. 1999). This discrepancy between pedigree and
molecular markers based on GS estimates may be
the result of assumptions that were made in the
calculation of genetic relatedness based on pedi-
gree. These estimates may be biased due to selec-
tion pressure, unequal parental contribution, and
also the relatedness of ancestors without a known
pedigree. Bohn et al. (1999) emphasised that bias is
generated by the breeding method employed on
wheat and barley where segregants are selected
for phenotypic similarity to the superior parent.
However, DNA markers have the advantage of
directly detecting sequence variation among culti-
vars and therefore, are not influenced by these
assumptions that are inherent to pedigree analysis
(Manifesto et al. 2001; Soleimani et al. 2002).
The present study is an investigation of DNA
markers for the characterisation of Tunisian
wheat germplasm. Cultivation of modern cultivars
with limited genetic diversity, over large areas of
Tunisia could result in a loss of the local genetic
diversity as well as increasing uniformity in the elite
germplasm. On the other hand, old durum wheat
cultivars known to possess characters for adapta-
tion to North African continental dryland dis-
played a high genetic diversity and were shown to
be genetically distant from modern cultivars.
Breeding programs should use the old cultivars
for hybridisation in order to obtain heterotic
hybrids as well for the conservation of adaptive
and resistant traits within Tunisian germplasm.
wheat cultivars should be followed by an agro-
nomic and morphologic evaluation in order to
develop DNA markers linked to adaptation traits
to biotic and abiotic conditions.
We are thankful to Dr M.S. Gharbi and M. Kasri,
Tunisian wheat breeders for providing durum
wheat cultivars and the two Aegilops species. This
workis partof the
‘Development of Biotechnological Research in the
Arab States’ with ICARDA. Financial support was
provided by the Arab Fund for Economic and
Akagi H., Yokozeki Y., Inagaki A. and Fugimura T. 1997.
Highly polymorphic microsatellites of rice consist of
AT repeats, and a classification of closely related cultivars
with these microsatellite loci. Theor. Appl. Genet. 94:
Autrique E., Nachit M.M., Monneveux P., Tanksley S.D. and
Sorrells M.E. 1996. Genetic diversity in durum wheat based
on RFLP, morphophysiological traits and coefficient of par-
entage. Crop Sci. 36: 735–742.
Barrett B.A. and Kidwell K.K. 1998. AFLP-based genetic diver-
sity assessment among wheat cultivars from the pacific
Northwest. Crop Sci. 38: 1261–1271.
Bœuf F. 1932. Le bl? e e en Tunisie. ‘La plante, le milieu physico-
chimique’. Revue de l’INRAT 1: 4–67.
Bohn M., Utz H.F. and Melchinger A.E. 1999. Genetic simila-
rities among winter wheat cultivars determined on the basis of
RFLPs, AFLPs, and SSRs and their use for predicting pro-
geny variance. Crop Sci. 39: 228–237.
Cox T.S., Lookhart G.L., Walker D.E., Harrell L.G.,
Albers L.D. and Rodgers D.M. 1985. Genetic relationships
among hard red winter wheat cultivars as evaluated by pedi-
gree analysis and gliadin polyacrylamide gel electrophoretic
patterns. Crop Sci. 25: 1058–1063.
Degha€ ? ?s M., Gharbi M.S., Faleh M., Zarkouna T. and
Chakroun M. 1998. Les acquis de l’am? e elioration g? e en? e etique
des c? e er? e eales en Tunisie. Actes du 3emeS? e eminaire International
sur la ‘Technologie’. Centenaire de l’INAT.
DograrN., Akin-YalinS.and AkkayaM.S.2000.Discriminating
durum wheat cultivars using highly polymorphic simple
sequence repeat DNA markers. Plant Breed. 119: 360–362.
Donini P., Law J.R., Koebner R.M.D., Reeves J.C. and
Cooke R.J. 2000. Temporal trends in the diversity of UK
wheat. Theor. Appl. Genet. 100: 912–917.
Eujayl I., Sorrels M.E., Baum M., Wolters P. and Powell W.
2002. Isolation of EST-derived microsatellite markers for
genotyping the A and B genomes of wheat. Theor. Appl.
Genet. 104: 399–407.
Guadagnuolo R., Savova Bianchi D. and Felber F. 2001.
Specific genetic markers for wheat, spelt, and four wild rela-
tives: comparison of isozymes, RAPDs, and wheat micro-
satellites. Genome 44: 610–621.
Incirli A. and Akkaya M.S. 2001. Assessment of genetic rela-
tionships in durum wheat cultivars using AFLP markers.
Genet. Resour. Crop Evol. 48: 233–238.
Jak? s se J., Kindlhofer K. and Javornik B. 2001. Assessment of
genetic variation and differentiation of hope genotypes by
microsatellite and AFLP markers. Genome 44: 773–782.
Karmous C., Slim-Amara H. and Daaloul A. 2002. Etude de
quelques crit? e eres physiologiques de s? e election in vitro pour la
tol? e erance a ` la salinit? e e chez le bl? e e dur (Triticum durum Desf.).
Revue de l’INAT 17: 18–24.
Kempthorne O. 1969. An Introduction to Genetic Statistics.
Iowa State University Press, Ames, Iowa.
Kendall M.G. and Stuart A. 1979. The Advanced Theory of
Statistics, vol. 2. 3rd edn. Griffin, London.
L€ u ubberstedt T., Melchinger A.E., Dußle C., Vuylsteke M. and
Kuiper M. 2000. Relationships among early European maize
inbreds: IV. Genetic diversity revealed with AFLP markers
and comparison with RFLP, RAPD, and pedigree data. Crop
Sci. 40: 783–791.
Manifesto M.M., Schlatter A.R., Hopp H.E., Sua ´rez E.Y. and
Dubcovsky J. 2001. Quantitative evaluation of genetic
diversity in wheat germplasm using molecular markers.
Crop Sci. 41: 682–690.
Mantel N. 1967. The detection of disease clustering and a gen-
eralized regression approach. Cancer Res. 27: 209–220.
Morgante M., Hanafey M. and Powell W. 2002. Microsatellites
are preferentially associated with nonrepetitive DNA in plant
genomes. Nat. Genet. 30: 194–200.
Nei M. and Li W.H. 1979. Mathematical model for studying
genetic variation in terms of restriction endonucleases. Proc.
Natl. Acad. Sci. USA 79: 5269–5273.
Pillen K., Binder A., Kreuzkam B., Ramsay L., Waugh R.,
Fo ¨rster J. and L? e eon J. 2000. Mapping new EMBL-derived
barley microsatellites and their use in differentiating German
barley cultivars. Theor. Appl. Genet. 101: 652–660.
Plaschke J., Ganal M.W. and Ro ¨der M.S. 1995. Detection of
genetic diversity in closely related bread wheat using micro-
satellite markers. Theor. Appl. Genet. 91: 1001–1007.
Powell W., Morgante M., Andre C., Hanafey M., Vogel J.,
Tingey S. and Rafalski A. 1996. The comparison of RFLP,
RAPD, AFLP and SSR (microsatellite) markers for germ-
plasm analysis. Mol. Breed. 2: 225–238.
Pujar S., Tamhankar S.A., Rao V.S. and Gupta V.S. 1999.
Arbitrarily primed-PCR based diversity assessment reflects
hierarchical groupings of Indian tetraploid wheat genotypes.
Theor. Appl. Genet. 99: 868–876.
Rezgui S., Yahyaoui A., Daaloul A., Slim-Amara H., Nachit M.
and Amri A. 1998. Relative performance of selected durum
wheat in the semi-arid regions. Proceedings of the Interna-
tional Wheat Genetic Symposium. Sasketchewan, Canada,
vol. 2, pp. 326–328.
Rohlf F.J. 2000. NTSYS-pc: Numerical taxonomy and multi-
variate analysis system. Version 2.1, Exeter Publications,
New York, USA.
Ro ¨der M.S., Plaschke J., Ko ¨nig S.U., Bo ¨rner A., Sorrells M.E.,
Tanksley S.D. and Ganal M.W. 1995. Abundance, variability
and chromosomal location of microsatellites in wheat. Mol.
Gen. Genet. 246: 327–333.
Ro ¨der M.S., Korzun V., Wendehake K., Plaschke J., Tixier M.H.,
Leroy P. and Ganal M.W. 1998. A microsatellite map of
wheat. Genetics 149: 2007–2023.
Russell J.R., Fuller J.D., Macaulay M., Hatz B.G., Jahoor A.,
Powell W. and Waugh R. 1997. Direct comparison of levels of
AFLPs, SSRs and RAPDs. Theor. Appl. Genet. 95: 714–722.
Saghai-Maroof M.A., Soliman K.M., Jorgensen R.A. and
AllardR.W.1984. RibosomalDNA spacer length
polymorphisms in barley: Mendelian inheritance, chromoso-
mal location,and populationdynamics.Proc. Natl. Acad.Sci.
USA 81: 8014–8019.
Allard R.W. 1994. Extraordinarily polymorphic microsatel-
lite DNA in barley: species diveristy, chromosomal locations,
and population dynamics. Proc. Natl. Acad. Sci. USA 91:
Soleimani V.D., Baum B.R. and Johnson D.A. 2002. AFLP and
pedigree-based genetic diversity estimates in modern cultivars
of durum wheat [Triticum turgidum L. subsp. durum (Desf.)
Husn.]. Theor. Appl. Genet. 104: 350–357.
Sourdille P., Tavaud M., Charmet G. and Bernard M. 2001.
Transferability of wheat microsatellites to diploid Triticeae
speciescarrying the A, B and D genomes.Theor. Appl. Genet.
Stachel M., Lelley T., Grausgruber H. and Vollmann J. 2000.
Application of microsatellites in wheat (Triticum aestivum
L.) for studying genetic differentiation caused by selection
for adaptationand use.
Van de Peer Y. and De Wachter R. 1994. TREECON for
Windows: a software package for the construction and draw-
ing of evolutionary trees for the Microsoft Windows environ-
ment. Comput. Appl. Biosci. 10: 569–570.