Introduction
Horse gram (Macrotyloma uniflorum (Lam.) Verdc) is one of the cool season food grain legume crops cultivated across 2 million ha in India, as well being cultivated in Myanmar, Nepal, Malaysia, Mauritius and Sri Lanka for food purposes, and in Australia and Africa primarily for fodder purposes. It is known for its resilience to abiotic stresses such as drought and salinity. It can also be grown in nutrient poor soils (Bhartiya et al., Reference Bhartiya, Aditya and Kant2015). Being a self-pollinated crop (Halder et al., Reference Halder, Datta, Mandal and Ghosh2012), pure-lines are the only cultivar options in horse gram. Pedigree selection of desirable recombinant inbred lines (RILs) for use as pure-line cultivars is the most widely used breeding method in horse gram. Very often, a crop breeder is confronted with the task of selecting a few among a large number of bi-parental/multi-parental crosses-derived segregating populations to implement pedigree selection to identify superior/transgressive RILs for use as pure-line cultivars. Early elimination of poor crosses based on an objective criterion enable increased allocation of resources to a few promising crosses for identifying transgressive RILs (Witcombe et al., Reference Witcombe, Gyawali, Subedi, Virk and Joshi2013; Bernardo, Reference Bernardo2020). Early generation (F2:3) prediction of frequency of transgressive RILs that could be derived in advanced segregating generations is one such objective criterion (Bernardo, Reference Bernardo2020). The objective of the present study is to predict the frequency of transgressive RILs and minimum population size required for their recovery from two crosses.
Experimental
The basic genetic material consisted of 259 and 245 F2 plants and a random sample of 39 and 25 F2:3 families derived from two crosses, namely HPKM320 (determinate) × CRIDA18-R (indeterminate) and IC 361290 (determinate) × Palem 1 (indeterminate) differing for growth habit (online Supplementary Table S1). The seeds of four parents and the two check varieties, namely PHG-9 and BGM-1, and F2:3 families were planted in a single row of 3 m length in randomized complete block design using two replications and those of individual F2 plants during 2020 rainy season in experimental plots of the University of Agricultural Sciences, Bangalore, India. The seeds of F2 plants were planted 0.2 m apart. Fifteen days after planting, seedlings of four parents, two check varieties and F2:3 families were thinned to maintain a spacing of 0.2 m between the plants to accommodate 12 plants and 0.3 m between the rows. The entire recommended production package was practiced to raise four parents, check varieties, F2 and F2:3 generations.
Data were recorded on 10 randomly selected plants in four parents, two check varieties and F2:3 progenies in each of the two replications, and all the F2 plants for four traits, namely, number of primary branches and pods and weights of sun-dried pods and grains. The average of these traits across 10 sample plants in each replication was computed and expressed as primary branches per plant, pods per plant, pod weight per plant (g) and grain weight per plant (g).
The predictors of transgressive RILs, namely (i) mid-parental value (m) and additive genetic effects [a] (online Supplementary Table S2) were estimated based on replication-wise traits' mean data by regressing coefficients of m and [a] of the means of four parents, F2 and F2:3 populations on to their observed traits' phenotypic values using the multiple regression model (Kearsey and Pooni, Reference Kearsey and Pooni1996) implemented in SPSS software version 16.0. The σ 2A was estimated as 2 × [(Mean squares MS) due to ‘between F2:3 families’ − MS due to error)/number of replications] (van Ooijen, Reference van Ooijen1989).
The frequency of recovering RILs that are likely to transgress the better parent, and the two check varieties, namely PHG 9 and BGM 1, was predicted as standard normal distribution integrals corresponding to the values of [a]/σ A, and (mean of the check variety − m)/σ A, respectively for each trait; where, ‘σ A’ is the standard deviation of additive genetic effects (Jinks and Pooni, Reference Jinks and Pooni1976, Reference Jinks and Pooni1980). The minimum population size required to guarantee (95%) that RILs transgress the pre-determined standards was predicted as the number (n) of RILs need to be raised such that probability of RILs that do not transgress pre-determined limits is less than 5% (Kearsey and Pooni, Reference Kearsey and Pooni1996). This probability was translated as (1 − P)n ⩽ 0.05, where, P and (1 − P) are probabilities of RILs that transgress and those that do not transgress pre-determined standards, respectively. The equation was solved for ‘n’ by applying logarithm to both the sides and rearranging the terms as n ⩾ log 0.05/log(1 − P).
Discussion
Given that most of the variation between F2:3 families is contributed by σ 2A (van Ooijen, Reference van Ooijen1989), significant differences among means of F2:3 families derived from both the crosses suggest significance of σ 2A in the inheritance of for all the traits (Table 1).
Table 1. Analysis of variance of F2:3 progeny families for four quantitative traits in horse gram
* Significant at P = 0.05.
** Significant at P = 0.01.
The predicted frequencies of RILs that transgressed the better parent and both the checks were higher from HPKM 320 × CRIDA 18-R than those from IC 361290 × Palem 1 for primary branches per plant (Table 2). On the other hand, the predicted frequencies of RILs that transgressed the better parent and both the checks were higher from IC 361290 × Palem 1 than those from HPKM 320 × CRIDA 18-R for pod weight per plant and grain weight per plant. For pods per plant, the predicted frequencies of RILs that transgressed the better parent and the check, BGM 1 were higher from IC 361290 × Palem 1 than those from HPKM 320 × CRIDA 18-R. Thus, our results suggest frequencies of transgressive RILs predicted from IC 361290 × Palem 1 were higher than those predicted from HPKM 320 × CRIDA 18-R for at least three of the four traits. Higher frequency of transgressive RILs predicted from IC 361290 × Palem 1 suggest greater dispersion of desirable alleles between the parents of this cross than those between the parents of HPKM 320 × CRIDA 18-R for all the traits except primary branches per plant. As expected, the minimum population size required to recover the transgressive RILs predicted from IC 361290 × Palem 1 was relatively smaller than those predicted from HPKM 320 × CRIDA 18-R. Our results clearly indicate better breeding potential of IC 361290 × Palem 1 than that of HPKM 320 × CRIDA 18-R in terms of predicted frequency of transgressive RILs. We believe that the objective criterion used in our study could be deployed by breeders of other self-pollinated crops to select a few promising crosses and identify transgressive RILs from them.
Table 2. Predicted frequency of RILs which transgressed the limits of means of parental and check varieties and minimum population size required for their recovery in horse gram
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1479262122000077
Acknowledgements
The senior author gratefully acknowledge the Department of Science and Technology (DST), Government of India for providing financial support in the form of INSPIRE-fellowship DST/INSPIRE Fellowship/IF180603 dated: 25/09/2019 for conducting thesis research for partial fulfilment for the award of Ph.D. degree by the University of Agricultural Sciences, Bangalore, India.
