Journal of Heredity Advance Access originally published online on January 4, 2006
Journal of Heredity 2006 97(1):55-61; doi:10.1093/jhered/esj009
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Inheritance of Time to Flowering in Chickpea in a Short-Season Temperate Environment
From the Department of Plant Sciences, University of Saskatchewan, Saskatoon, SK, Canada S7N 5A8
Address correspondence to Yadeta Anbessa at the address above, or e-mail: yadeta.kabeta{at}usask.ca.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Time to flowering is central in determining the adaptation and productivity of chickpea in short-season temperate environments. We studied the genetic control of this trait in three crosses, 272-2 x CDC Anna, 298T-9 x CDC Anna, and 298T-9 x CDC Frontier. From each cross, 180 F2 plants and parents were evaluated for time to flowering under greenhouse conditions. In summer 2004, multiple generations including P1, F1, P2, F2, and F2:3 (also called MG5) were evaluated for time to flowering under field conditions. The data on time to flowering in the F2 populations were continuous in distribution but deviated from normal distribution. The F2:3 families derived from this showed a bimodal distribution for time to flowering, a typical case of major-gene inheritance model with duplicate recessive epistasis. A joint segregation analysis of MG5 also revealed that time to flowering in chickpea was controlled by two major genes along with other polygenes. Late flowering was dominant over early flowering for both major genes with digenic interaction between them, mainly an additive x additive type. This information can be used to formulate the most efficient breeding strategy for improvement of time to flowering in chickpea in short-season temperate environments.
Chickpea (Cicer arietinum L.) crops often experience short growing seasons as a result of drought, heat, or end-of-season frost (Khanna-Chopra and Sinha 1987). Early flowering is a key factor in the formation and maturation of pods before the occurrence of these abiotic stresses. Kumar and Abbo (2001) have reported that time to flowering plays a central role in determining the adaptation and productivity of this crop in short growing environments.
The flowering time of chickpea is variable depending on season, sowing date, latitude, and altitude (Summerfield and Roberts 1988). According to Roberts et al. (1985), time to flowering was a function of temperature and photoperiod in chickpea. Ellis et al. (1994) further noticed that in some chickpea genotypes, time to flowering was influenced by photoperiod and temperature, whereas in others, flowering time was determined solely by photoperiod.
Gumber and Sarvjeet (1996) studied the genetics of time to flowering in three crosses of chickpea and found that it was controlled by two genes. Kumar and van Rheenen (2000) observed a bimodal distribution for time to flowering in chickpea and deduced the presence of one major gene (Efl-1/efl-1) plus polygenes for this trait. Or et al. (1999) also supported this result, but they associated the major gene with sensitivity to photoperiod (Ppd/ppd). Kumar and Abbo (2001) suggested that the major early-flowering alleles efl-1 and ppd may be located at the same locus, although no experimental evidence supporting this hypothesis is yet available. Analysis of quantitative data by Cho et al. (2002) revealed a quantitative trait locus (QTL) for days to 50% flowering. But the distinction between a major gene and a QTL is sometimes rather artificial because once a QTL is identified and located, it effectively becomes a major gene (Knott et al. 1991).
The genetics of time to flowering needs to be sufficiently understood in order to fine-tune cultivars to the demands of a particular environment. The ability to efficiently manipulate time to flowering is a crucial component of chickpea improvement (Kumar and Abbo 2001). The above evidence indicates that genetic variation for time to flowering is mediated by genes of variable, rather than equal, effects. If a major gene with a significant effect on the variation can be identified, this can be manipulated in a directed fashion. However, cosegregating polygenes and environmental effects make the detection of major genes difficult.
Joint segregation analysis (JSA) was applied for the analysis of major gene and polygenes mixed quantitative variation in plants in recent years (Wang and Gai 2001). It has been used to analyze mixed-inheritance models in human and animal populations over the last four decades (Elston and Steward 1973; Knott et al. 1991). In brief, the method works as follows. First, some possible genetic models are hypothesized and likelihood functions are established for the different genetic models. Then, maximum likelihood estimates of the parameters contained in each genetic model are obtained. The best-fitting genetic model is selected based on Akaike's information criterion (AIC) (Akaike 1977 cf. Knott et al. 1991). The objective of this study was to determine the most appropriate genetic model describing the variation in time to flowering in chickpea in a short-season temperate environment.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Three crosses, 272-2 x CDC Anna, 298T-9 x CDC Anna, and 298T-9 x CDC Frontier, were made at the University of Saskatchewan, Saskatoon, Canada. 272-2 and 298T-9 were selected from field nurseries as early-flowering lines (Table 1). They both had ICCV 96029 as one of their parents, which was reported as the world's earliest flowering chickpea germplasm to date (Kumar and Rao 2001). Based on ratings in Saskatchewan, CDC Anna and CDC Frontier are late-flowering cultivars of desi and kabuli market classes, respectively. The F1 was advanced to F2, and 180 F2 plants from each cross were evaluated in the greenhouse in summer 2003. A single seed was taken from each F2 plant to produce F3 plants in the same greenhouse during fall 2003/winter 2004. In both cases, one plant was grown in each 20-cm-diam pot filled with Redi-Earth soil (WR Grace and Company, Ajax, ON, Canada). Photoperiod was maintained at 18 h, and mean air temperature was 24°C ± 3°C in the greenhouse. During the summer 2004, all the P1, F1, P2, F2, and F2:3 populations from saved seed at each generation were evaluated in a field experiment near Saskatoon (lat 52°09'N, long 106°36'W). The maximum daylength at this location was about 18 h. Fifteen space-planted individual plants were used for each P1, F1, and P2, whereas the F2 individuals ranged from 121 to 143 per cross. The F2:3 generation had 115 families of about 30 plants each in all the three crosses. Seasonal temperature in 2004 at this location is shown in Figure 1. Time to flowering, that is, the number of days from planting to flowering was recorded. The distribution patterns of time to flowering data for F2 populations and their F2:3 progeny were analyzed.
|
|
Joint Segregation Analysis
Genetic Models
Five classes of genetic models were considered to select the one that best explains the variation in time to flowering in chickpea (Table 2). Taking into account gene action (additive, dominance, additive-dominance, or additive-dominance-epistasis), we further set up model types within each class, and overall 26 scenarios were considered.
|
Estimation of Component Parameters
The maximum likelihood estimates of the component parameters in each genetic model were obtained by the expectation and iterated maximization algorithm (Zhang et al. 2003). Suppose a quantitative trait is controlled by one major gene AA and polygenes. The F1 from a cross between high and low parents would be Aa for the major gene. Three genotypes are possible at F2 with segregation ratio of 1:2:1. The F2:3 populations will have the same mixture of genotypes as F2, but the proportion of individuals will change. Because of the effect of polygenes and environmental variance, for any given major genotype, the phenotypes of all individuals are independently and normally distributed and therefore the distribution of MG5 would be:
 |
 |
 |
 |
P1, F1, and P2 are assumed to have equal variance (
2) with their respective means of µ1, µ2, and µ3; µ41, µ42, and µ43 are means of the three major F2 genotypes AA, Aa, and aa, respectively; 
(common variance of components in F2); µ51, µ52, and µ53 represent means of F2:3 families derived from AA, Aa, and aa, respectively; 
is the variance of the component having mean µ52; and 
and 
are the common variance of the nonsegregating F2:3 families for the locus. Accordingly, the component parameters estimated include µ1, µ2, µ3; µ41, µ42, µ43; µ51, µ52, µ53; and 2, 
Model Selection
To allow for different hypotheses depending on different numbers of unknown parameters, the hypothesis that maximizes the expected entropy is chosen (Akaike 1977 cf. Knott et al. 1991). For this purpose, we chose the hypothesis that leads to the smallest AIC as the best fitting:
 |
Estimation of Genetic Parameters
Once the component parameters are set, it is possible to derive genetic parameters from them. Considering the above example again, we obtain the following relationships for major gene and polygenes mixed inheritance:
 |
 |
 |
 |
 |
 |
 |
 |
 |
Variances were partitioned into components based on the following relationships:
where 
is the common variance across all F2 genotypes, 
is the polygenic variance in F2 population, and 2 is the environmental variance.
where 
and Vmg are variances due to polygenes and major gene in F2:3 population, respectively, and n is the number of plants observed. 
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Frequency Distribution of Time to Flowering
The F2 populations evaluated in the greenhouse showed continuous variation for time to flowering (Figure 2). The majority of the individuals fell between the two parents for time to flowering, but a few individuals were 1 to 2 days earlier than the early parent and others were up to 3 days later than the respective late parent. The data on time to flowering in this F2 population deviated from normal distribution for all three crosses (P
.03), as revealed by the Shapiro-Wilk's test of the SAS proc univariate normal (SAS Institute Inc., 1999). The distribution of time to flowering data in these F2 populations was skewed toward the late parental type (Figure 2).When F2 was advanced to F3, some genotypes derived from the late-flowering F2 plants flowered earlier than the population mean at F3, whereas a few others, which were early at F2, flowered later than the population mean at F3 (Figure 3). Late flowering is dominant over earliness in chickpea (Gumber and Sarvjeet 1996; Kumar and van Rheenen 2000; Or et al. 1999) and F2 genotypes, which were late to flower, could be early at F3 because of segregation in the heterozygous plants. The opposite move from early in F2 to late in F3 would probably indicate the involvement of epistatic gene action. The early and late parents included as checks were earlier and later than the population mean in time to flowering across both tests, respectively.
|
|
Summer 2004 was cooler than average at Saskatoon (Figure 1). As a result, crop growth was slower than average and flowering was delayed for chickpea. Under these conditions, the early parents flowered in 52–53 days from planting, whereas the late parents took 60–61 days (Figure 4). The distributions of time to flowering data for the F2:3 families evaluated in the field were also continuous, but these followed a bimodal pattern (Figure 4). Classification of the time to flowering data into early (55 days or earlier) and late (later than 55 days) matched a segregation ratio of late to early flowering of 9:7 (
2 = 0.57, P = .45 for 272-2 x CDC Anna; 2 = 0.89, P = .35 for 298T-9 x CDC Anna; and 2 = 2.26, P = .13 for 298T-9 x CDC Frontier). This indicates that time to flowering was governed primarily by two genes with duplicate recessive epistasis between them in the short-season temperate environment.
|
Joint Segregation Analysis
The JSA revealed that the variation in time to flowering in chickpea was best explained by mixed two major genes plus polygenes model class (Table 3), confirming the evidence from the frequency distribution pattern of the F2:3 populations indicated above. Model E-1 had lowest AIC for the 272-2 x CDC Anna and 298T-9 x CDC Anna populations and model E-2 for the 298T-9 x CDC Frontier population. The difference is that in model E-1 there is interaction between the two major genes, but model E-2 assumes additive-dominant type of gene action. However, the C scaling test of Mather and Jinks (1982) revealed the presence of epistasis in all the three crosses (data not shown). The segregation ratio in F2:3 families also indicated the involvement of epistatic gene action.
|
The majority of the variation in time to flowering was accounted for by the two major genes (Table 4). The contribution of polygenes to the total phenotypic variation was very low. Heritability of major genes was high, greater than 60% across generations and crosses. For both major genes, the late-flowering alleles showed dominance over early-flowering alleles (Table 4). The JSA also revealed the presence of digenic interaction between the two major genes, mainly additive x additive type.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
This study revealed that time to flowering in chickpea in high-latitude, cool-season environments followed a two major genes plus polygenes mixed-inheritance model. The two major genes determined the majority of the phenotypic variation (>65%) for this trait, and the contribution of polygenes was minimal. Previous reports on the inheritance of time to flowering in chickpea were inconsistent, and all came from short-day, warm-temperature environments. In a preliminary report from India, based on crosses among the early (ICCV 2) and two late (GL769, BG276) parents, Gumber and Sarvjeet (1996) proposed that time to flowering was controlled by duplicate genes. However, using the same early-flowering parent ICCV 2, Kumar and van Rheenen (2000) found a single major gene plus polygenes mode of inheritance for time to flowering. Or et al. (1999) also reported a single major gene for time to flowering, supporting the latter finding.
ICCV 96029, which was an indirect source of early-flowering alleles for our populations (see Materials and Methods), was developed from a cross of two early-flowering genotypes ICCV 2 x ICCV 93929 (Kumar and Rao 2001). It was about 1 week earlier than either of the parents at the International Crops Research Institute for the Semi-Arid Tropics, Patancheru, India (Kumar and Rao 2001). This genotype likely has additional alleles for early flowering, which strongly supports our finding. The duration from sowing to flowering in other legumes such as common bean (Phaseolus vulgaris) and pigeonpea (Cajanus cajan) is also under the control of two genes (e.g., Craufurd et al. 2001; Kornegay et al. 1993).
Time to flowering is determined by three factors: response to photoperiod (usually the most important factor), response to temperature, and "earliness per se" genes (Snape et al. 2001). We did not have sufficient information from this research to determine to which group the two major genes detected in this study belong. But the differences in response of parental genotypes to changes in growing conditions implied interaction with the environment (photoperiod and temperature). Physiological studies revealed that time to flowering is a function of temperature and photoperiod in chickpea (Roberts et al. 1985) and the two major genes detected may each respond to either factor. Snape et al. (2001) also reported that different major genes controlled temperature and photoperiod effects on time to flowering and that earliness per se is generally considered a QTL in wheat.
Estimates of genetic parameters provide an indication of the relative importance of the different types of gene effects affecting the total genetic variation (Hayman 1958). In this study, epistatic gene effects were present in sufficient magnitude to be considered important. The estimates of the additive x additive gene effects have greater relative magnitude than the other two types of digenic epistasis (additive x dominance and dominance x dominance) for this trait. Additive x additive epistasis is generally fixable but requires delayed or later generation selection. Arshad et al. (2003) also noticed epistatic gene effects for time to flowering in chickpea. However, Malhotra and Singh (1989) did not observe epistasis for any of the important agronomic traits including days to flowering. This was probably due to the differences in allelic constituents of the parental genotypes used.
The detection of major genes for time to flowering in chickpea demonstrates that this trait can easily be incorporated into the desired genetic background. The backcrossing or single-seed descent breeding methods could effectively be deployed to advance time to flowering in chickpea. Our results showed that incorporation of these two alleles advanced flowering date by about 1 week under western Canadian conditions at lat 52°N. This effect would likely be greater under short-day environments at lower latitudes, where flowering is greatly delayed for late (more photoperiod sensitive) genotypes. Hence, these two alleles play an important role in accelerating flowering time in chickpea.
Chickpea is a highly indeterminate species and early flowering may extend the duration of the reproductive period (Subbarao et al. 1995). In short-season temperate environments, the duration of reproductive period is determined by the commencement of flowering and the end-of-season drought or frost that terminates seed setting and growth. A longer reproductive period, brought about by early-flowering alleles, could enhance seed yield in chickpea by allowing formation of a relatively large number of pods and through longer grain-filling duration (Or et al. 1999). Therefore, more progress could be made with respect to yield and earliness by incorporating the two early-flowering alleles reported herein into adapted genetic backgrounds.
|
Acknowledgments |
|---|
This project was supported by a grant from Saskatchewan Agriculture, Food and Rural Revitalization. We also acknowledge the financial support provided by the Saskatchewan Pulse Growers. Dr. Y. M. Zhang (University of California, Davies) kindly provided the JSA program and also helped with the data analysis.
|
Footnotes |
|---|
Corresponding Editor: Reid Palmer
Received April 5, 2005
Accepted November 28, 2005
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
-
Akaike H, 1977. On entropy maximum principles. In: Applications of statistics (Krishnaiag G, ed.). Amsterdam, The Netherlands: North-Holland Publishing Company; 27–41.
Arshad M, Bakhsh A, Zubair M, and Ghafoor A, 2003. Genetic variability and correlation studies in chickpea (Cicer arietinum L.). Pak J Bot 35:605–611.
Cho S, Kumar J, Schultz JL, Anupama K, Tefera F, and Muehlbauer FJ, 2002. Mapping genes for double podding and other morphological traits in chickpea. Euphytica 128:285–292.[CrossRef]
Craufurd PQ, Soko HS, Jones JK, and Summerfield RJ, 2001. Inheritance of duration from sowing to first flowering in pigeonpea. Euphytica 119:323–333.[CrossRef]
Ellis RH, Lawn RJ, Summerfield RJ, Qi A, Roberts EH, Chays PM, Brouwer JB, Rose JL, Yeates SJ, and Sandover S, 1994. Towards the reliable prediction of time to flowering in six annual crops: V. Chickpea (Cicer arietinum). Exp Agric 30:271–282.
Elston RC and Steward J, 1973. The analysis of quantitative traits for simple genetic models from parental, F1 and backcross data. Genetics 73:695–711.
Gumber RK and Sarvjeet S, 1996. Genetics of flowering time in chickpea: a preliminary report. Crop Improv 23:295–296.
Hayman BI, 1958. The separation of epistatic from additive and dominance variation in generation means. Heredity 12:371–390.
Khanna-Chopra R and Sinha SK, 1987. Chickpea: physiological aspects of growth and yield. In: The chickpea (Saxena MC and Singh KB eds). Wallingford, UK: CAB International; 163–189.
Knott SA, Haley CS, and Thompson R, 1991. Methods of segregation analysis for animal breeding data: a comparison of power. Heredity 68:299–311.
Kornegay J, White JW, Dominguez JR, Tejada G, and Cajiao C, 1993. Inheritance of photoperiod response in Andean and Mesoamerican common bean. Crop Sci 33:977–984.
Kumar J and Abbo S, 2001. Genetics of flowering time in chickpea and its bearing on productivity in semiarid environments. Adv Agron 72:107–138.
Kumar J and Rao BV, 2001. Registration of ICCV 96029, a super early and double podded chickpea germplasm. Crop Sci 4:605–606.
Kumar J and van Rheenen HA, 2000. A major gene for time of flowering in chickpea. J Hered 91:67–68.
Malhotra RS and Singh KB, 1989. Detection of epistasis in chickpea. Euphytica 40:169–172.
Mather K and Jinks JL, 1982. Biometrical genetics, 3rd ed. London, UK: Chapman and Hall.
Or E, Hovav R, and Abbo S, 1999. A major gene for flowering time chickpea. Crop Sci 39:315–322.
Roberts EH, Hadley P, and Summerfield RJ, 1985. Effects of temperature and photoperiod on flowering in chickpeas (Cicer arietinum L.). Ann Bot 55:881–892.
SAS Institute Inc., 1999. SAS language and procedure: usage version 8-2. Cary, NC: SAS Institute Inc.
Snape JW, Butterworth K, Whitechruch E, and Worland AJ, 2001. Waiting for fine times: genetics of flowering time in wheat. Euphytica 119:185–190.[CrossRef][Web of Science]
Subbarao GV, Johansen C, Slinkard AE, Nageswara Rao RC, Saxena NP, and Chauhan YS, 1995. Strategies for improving drought resistance in grain legumes. Critic Rev Plant Sci 14:469–523.
Summerfield RJ and Roberts EH, 1988. Photothermal regulation of flowering in pea, lentil, faba bean and chickpea. In: World crops: cool season food legumes (Summerfield RJ, ed.). Dordecht, The Netherlands: Kluwer Academic Publishers; 911–922.
Wang JK and Gai JY, 2001. Mixed inheritance model for resistance to agromyzid bean fly (Melanagromyza sojae Zehntner) in soybean. Euphytica 122:9–18.[CrossRef]
Zhang YM, Gai JY, and Yang YH, 2003. The EIM algorithm in the joint segregation analysis of quantitative traits. Genet Res Camb 81:157–163.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||