|
Introduction
Quantitative
genetic study was initiated in 2000 to provide information on the
genetics of traits of agronomic interest. Three sets of parents, that
represent clones adapted to the three major agro-ecologies most
important for cassava in the tropics: low land semi-arid; acid savannas
and mid-altitude valleys, were crossed in a diallel fashion to produce F1
families used in the study. These were planted in September 2001 and
evaluated in May 2002. Griffing’s Method 4 (only crosses, no parents
and reciprocal crosses) (Griffing 1954) was used to estimate the General
Combining Ability (GCA) and Specific Combining Ability (SCA) effects and
their relative importance in the crosses. Data from each agronomic zone
was analyzed separately as it represents a different set of parents.
Percent Dry matter content (DMC), dry matter yield, harvest index (HI),
plant type (PT) and foliage weight were used in the analysis, because of
their importance on yield (Bryne, 1984). GCA and SCA estimates were
used for selecting parents and families to be used in the bulk
segregation analysis and QTL mapping studies respectively.
High positive and
negative values were estimated for both GCA
and SCA for all the traits.
Estimates varied for the different sites within location but tended to
follow a pattern, especially for percent dry matter content, dry yield
and harvest index. From the results emphasis was placed on percent dry
matter given its stability across environments within agro-ecologies for
the development of markers associated with QTLs for use in breeding.
Parents selected for the generation of larger sized progeny for QTL
analysis were planted in a crossing block in Palmira and crosses
generated. Using SCA 23 families with the high diversity for dry matter
were selected, of this two GM 312 and GM 313 were used to initiate Bulk
Segregant Analysis (BSA), and the rest planted in the field for another
two years to provide data to test the resultant markers.
In the field
experiments all agronomic traits were highly correlated to yield
confirming their importance to yield. Percent dry matter and the number
of tubers per plant were not significantly correlated to the number of
plants harvested suggesting that they are not easily influenced by
environment, implying that you do not need a large replicated trial with
many plants per genotype to determine percent dry matter. Significant
negative correlations were estimated between Cassava frog Skin Disease (CFSD)
and all the yield parameters. Highest estimate was obtained with
percentage dry matter (-0.33) followed by harvest index (-0.24), dry
tuber yield and number of commercial roots (-0.23), fresh tuber yield
and number of tubers (-0.20). The high negative correlation between DMC
and CFSD is of concern as it is likely to lead to false phenotypic data.
Two markers SSRY
150 (R2=18.1) and SSRY 160 (R2=29.3) showed association in family GM 313
with a value high enough for practical use in breeding. Screening the
markers on the selected families from the diallel cross revealed
association with DMC in other families that have SM 1741-1 as a parent,
revealing a strong genetic background-specific association between SSRY
160 with DMC. In another BSA experiment with an Inter-specific cross,
five markers SSRY 99 (R2=22.68), SSRY 141 (R2=35.89,
NS 169 (R2=20.01), SSRY 11 (R2=26.85) were
identified to be associated with favorable alleles and are useful
markers for the introgression of these alleles into cultivated gene
pools.
Eight sets of
crosses comprising of 2935 seeds were generated from the parents, of
these 1885 seedlings were established. DNA was extracted from the
largest families GM 901, CM 9953 and GM 252 for BSA analysis. In May
2005 the seedling nursery was harvested, evaluated for yield related
agronomic traits and cloned. Using DMC data bulks were made for BSA.
|
