Abstract:

—same (Sesamum indicum L.) belongs to the family Pedaliaceae in the order lamiales which
zmtains over 30 species but with Sesamum indicum as the only cultivated species. It is the most
mcient oilseeds crop known to mankind. It is normally called 'Simsinf in Eastern Africa and is
traditional and important oilseed crop in Uganda. It is often referred to by the epithet '''the
steen of oil seeds''' because of its nutritive value, quality and quantity of its oil which is rich in
?Lamin E and a significant amount of linoleic acid that can control blood cholestrol levels. Most
f the cultivated sesame varieties in Uganda are the local landraces which may be found mixed
nd are low yielding with the average national yield reported to be about 600 kg/ha. The local
andraces which are the main cultivars grown by farmers in Uganda are characterized by poor
-aits such as few capsules on the main stem zone, setting of the first capsule high up on the main
¦tern from the ground, and wide internode lengths. The overall objective of this study was to
ontribute to knowledge that willimprove on the sesame yield contributing traits through genetic
liversity of the germplasm in Uganda. The genetic diversity of Uganda sesame germplasm was
issessed from a collection of 1 15 accessions of sesame landraces and 85 breeding lines including
introductions using a set of newly developed microsatellite markers. The results of analysis of
nolecular variance (AMOVA) for landraces revealed that the majority of the variance occurred
imong the individuals accounting for 75% of the total variation and only 6% was attributed to
iifferences among the districts, suggesting a high gene flow (Nm = 4.476). These results were
supported by the PCoA analysis as well as the neighbor joining (NJ) tree which revealed no
Mustering of the accessions according to their geographic origin. The statistical analysis of 10
igromorphological traits indicated no clear pattern related to the geographic origin. Such a poor
’rouping was indicative of significant gene flow across geographic domains and could be
explained either by a high outcrossing rate and/or through extensive seed trading. Chloroplast-
SSR analysis using the marker ccmp2 revealed 3 haplotypes, with haplotype B being
jredominant and present in 95% of the individuals. The analysis of 9 nuclear SSR regions
showed mean Hc values (gene diversity. D) from 0 to 0.373 per line. The gene flow (Nm) among
he accessions grouped according to their countries of origin was 1.387 suggesting constant gene
low which was also supported by the observed outcrossing rates of between 3.1% - 62.5%. The
-\M0VA for breeding lines from different countries indicated that the majority of the variance
X
xvii
was occurring among the individuals accounting for 69% of the total variation and only I 1 % of
the variation was attributed to differences among the countries of origin. This was also supported
by the population structure analysis using the model-based approach revealing 4 groups (K = 4),
as well as the PCoA and the N.I tree analysis, where neither an origin nor an accession-associated
pattern was observed. However, the analysis of 10 agromorphological traits indicated a certain
conservation of some morphological traits, reflecting the geographic origin of the analysed lines.
For the inheritance and gene action in genotypes for yield traits, the dial lei mating design where
an 8x8 diallel cross mating design with 8 parents and 28F1 s showed high GCA effects were
recorded for days to flowering, height of first capsule and number of branches, branches per
plant and capsule length. Desirable SCA effects were recorded for Sesim2//5181 x Renner 1-3-1-
1 for reduced height of First branch, Sesiml x Sesim2 for number of branches per plant, capsules
on branches and yield per plant. High Baker’s ratio was recorded for most of the traits except for
yield per plant. CGDbs was high for most of the traits except for days to maturity, plant height,
capsules on main stem and capsule width. These results suggested that both additive and nonadditive
gene actions played a greater role on these traits. Sesim 1 had the highest number of
desirable traits scored for GCA effects and could be considered as a parent for crossing to
produce desirable progeny. Sesim I x Sesim 2, Sesim 1 x Ajimo A1-5 and Sesim2//5181
SPS1438-1-6-4 recorded positive significant SCA effects for yield per plant and therefore they
could be recommended for hybrid seed production. The line x tester mating design where nine
female local cultivars were used as lines and 5 male introductions were used as testers to produce
45 Fl hybrids which were evaluated together with the 14 parents in three replications revealed
that the mean squares across locations had high significant difference for all the traits recorded
indicating that locations had effect on those traits. Variances for the males were higher than the
variances for the females for all the traits except capsule width suggesting greater variability
among the males than the females. Combined general combining ability (GCA) across locations
showed that GCA for the lines and testers were significantly positive for a number of traits.
Specific combining ability (SCA) was only positively significant for days to flowering, number
of capsules on branches, number of capsules on main stem, total number of capsules and capsule
length. Baker’s ratio was high for most of the characters and medium for plant height, number of
capsules on branches, number of capsules on main stem and capsule length. Coefficients of
xviii
II
determination for narrow sense heritability (CGDns) were high for days to flowering, height of
first branch and height of first capsule. Among the lines, Sesim 2, Ajimo A1-5. EM 15-1-5,
Ajimo AI-6//7029, Local 158//6022 and (Y-1//Local 158)-l-2-l and among testers 4036-1-10-2
and SPS1438-1-6-4 were desirable for different traits.The results of analysis of genotype x
environment interaction to identify stable and high yielding genotypes showed that AMMI
analysis of variance for seed yield were highly significant (P<0.00l). The combined analysis of
variance indicated that season, season x location, genotype and location x genotype had
significant (P<0.001) variation. The GGE biplot suggested the existence of only one sesame
mega-environment with genotype G9 winning in that mega-environment. Genotypes were
categorized into stable and high yielding, stable but poor yielding, unstable but good yielding
and unstable and poor yielding.