| dc.description.abstract | Sorghum (Sorghum bicolor L. Moench) is an important cereal in global agriculture. Sorghum is a
leading cereal in arid and semi-arid agriculture, ranking fifth in importance among the world's
grain crops. Efforts in sorghum improvement of the crop for important agronomic and food traits
are increasingly adopting genetic engineering. The use of transgenic crops has been associated
with several potential risks in cropping systems including enhancement of invasive ability of
hybrids between crops and their wild progenitors. The study investigated pollen mediated
transfer of S bicolor alleles to wild sorghum populations (S sudanense; S halepense; S bicolor
ssp. verticilliflorum), described the population structure of wild sorghums in western Kenya and
described the effects of introgression on hybrid fitness and fecundity. The study also evaluated
the effectiveness of molecular ~approaches in detecting allelic transfer during interspecific
hybridization. Phenotypic methods, peR, diversity analysis and genetic distances were used to
evaluate controlled and spontaneous interspecific hybridization events and gene flow in the
greenhouse and field studies. Real time peR and Multiplex-Pf.R were evaluated for the
efficiency of detecting crop alleles in the wild sorghums growing in sympatry with cultivated
species. Replacement series were used to evaluate the relative competitive abilities of the F1
progenies in the presence or absence of Bt. larvicides. In the results, unique and null SSR alleles
in wild sorghums and high polymorphisms were observed among the accessions. Interspecific
hybridization was observed between S. halepense x S. bicolor and S. sudanense x S. bicolor.
There exists variation in interspecific hybridization among species within the sorghum genus.
There were significant phenotypic variations on vegetative traits with F1 progenies having
profuse tillering and branching than the parental populations. Phenotypic characteristics of the F1
varied with the wild parents involved in crosses. Microsatellite loci showed differential banding
XIX
patterns between parents and the FI progeny. The crosses had midway clusters between
respective parents on dendogram analysis. The real time-PCR technique was effective in
qualitative, quantitative and genotyping analysis of crop alleles in both crop and wild
backgrounds. Melting point analysis was valuable in determination and discrimination of SSR
alleles. Both crossing point analysis and melting point analysis were valuable in differentiating
wild sorghum populations from the different agro-ecological zones from Western Kenya. Crop
alleles were observed in wild, sorghum populations in agro-ecological zones in Homa Bay, Siaya
and Busia counties of Kenya. In addition wild sorghum population had moderate to high
diversity on SSR loci assayed. The population had low inbreeding, low genetic differentiation
and low to moderate deviation from HW equilibrium. Intra-populations diversity (Hs) was larger
than inter-population diversity (OST) in all populations. This showed pollen and seed mediated
gene flow was important in the wild sorghum populations and disqualified any recent population
bottlenecks in the wild populations. Wind parameters showed significant effect in pollen
mediated gene flow in concentric plots. There were significant species effects in the flow of
genes in UFD plots. S. bicolor ssp. verticilliflorum had low gene flow frequencies in most
directions relative to S. halepense and S. sudanense. More interspecific hybridization events
were observed above lOOm from pollen source in S. halepense and S. sudanense and a regression
model LoglOY= LOglOA+ Bx best explained the short distance and long distance flow of crop
genes in the open pollinated concentric design plots. Short distance gene flow was recorded 5m
from the donor plot while long distance 'jump" dispersal was observed at between 80-100m. The
maximum distance of gene flow in the plots was seen at 200m. Interspecific hybridization among
sorghum species was observed using Multiplex-PCR. Template concentration, primer quality and
reaction conditions were important in efficient amplification and analysis of multiple loci. Using
xx
the technique 2 to 3 S. bicolor loci were amplified concurrently in the wild species, with
substantial reduction in peR costs. S. sudanense had higher affinity towards hybridization with
the crop as well as the wild materials showing "bridge species" features. F1 progeny obtained
from the interspecific crosses showed heterosis on vegetative morphological parameters. Fitness
associated traits of F1 hybrids grown in competition with their parents were enhanced and varied
among species. The F1 progenies had equal or lesser ratoons than the wild parents, but they had
more seed as compared to their parents. All F1 populations had high levels of seed dormancy and
poor germination. Presence of Bt larvicides increased seed production in parental as well as the
F 1 populations grown in competition at different parent:F 1 ratios. This increase was seen in all
competition plant mixtures in replacement series except where the F1 was grown in 25% ratios.
There were more panicles and more seeds in the Bt sprayed populations. There exists variation in
interspecific hybridization within the sorghum genus giving rise to high diversity in wild
sorghum populations. The F1 hybrids had high heterosis on both vegetative and reproductive
features but a fitness girdle was observed with poor germination of the F1. In order to analyze the
effect of transgenes in wild sorghums, there is need to evaluate the specific events to examine
their fitness contribution in the progeny. The effects of interspecific hybridization can be
minimized by the use of molecular tools such as peR, real time-Pf'R and multiplex-Pf'R for
rapid tracking of crop genes in wild sorghums. In addition physical and genetic confinement
tools can be utilized for ease of integrating transgenics in cropping systems. | en_US |
