Findings of the present study confirm our hypothesis that the use of within-breed selection strategy improves egg and meat production of IC in Rwanda. Furthermore, this study revealed that based on rates of genetic gain and inbreeding, GBS outperformed CBS. This was confirmed by a low and high response to selection for overall breeding goal and individual traits realised in CBS and GBS, respectively.
4.1. Rates of Genetic Gain and Inbreeding of Indigenous Chicken in CBS and GBS
Generally, the rate of genetic gain for the breeding goal of IC in CBS in this study was adequate, but inferior to that realised in GBS (Table 2), confirming the findings from the previous studies [72, 51, 18]. This could be attributed to an increased accuracy of index realised in GBS (Table 2). High accuracy has been associated with high selection response [18]. It indicates, therefore, that GBS would increase the values of genetic improvement. The GBS is associated with higher accuracy than CBS because genomic relationships are more accurate than pedigree-based relationships [46]. Accuracy is positively correlated with heritability [79] and in GBS, the marker information is entirely heritable [65, 15]. This is because markers are linked to a gene affecting the trait [46]. Besides, the increase in accuracy is attributed to the upgraded measurement of the relationships between animals and better prediction of the Mendelian sampling terms [11, 28]. With GBS, Mendelian sampling terms are better exploited than with CBS ([10]), and thus accuracy from GBS increases compared to those realised in CBS. This is because GBS captures not only the additive genetic relationship between individuals but also the information about linkage disequilibrium between markers and traits [35]. This study suggests GBS be a breeding scheme that results in more rapid genetic progress compared to CBS.
The simulated performance of the chickens selected in the nucleus using CBS was less than the performance of their descendants raised in the smallholder farms (Table 2), confirming G × E [53, 51]. Since, in practice, environments in which chickens are kept and selection is done, are often different, the breeding goal should, however, reflect the economic and production environment in which the animals are reared [60, 44]. This is done by involving genetic correlations caused by G × E [74], which is generally lower than unity [17]. A loss in genetic gain should be expected when G × E is less than 0.8 [53, 21]. This is because, G × E reduces the accuracy of selection and, hence, effective heritability of EBV for commercial population performance [16]. The G x E is a cause of genetic variation with regards to the environment [51] as genes are affected by environmental change [70]. The effect of selection is to change gene abundance [27] and adaptation of animals in a new environment can, therefore, happen through the modification of their chromatin structure, especially change in the gene sequence through recombination, genetic drift and mutation, [70]. Improvement obtained in the nucleus by using CBS would not, therefore, be fully realised in the production environment where G x E interaction is significant. These findings are supported by [14, 21, 51] who demonstrated that using CBS, the performance of selected animals in the nucleus could be a poor predictor of performance of their offspring under field conditions due to G x E. This study advises that by using CBS, G x E should, therefore, be considered to develop superior germplasm among IC that performs optimally under specific management practices.
This study revealed, however, that by using GBS, the performance of chickens in the nucleus was not different from that of their descendants in smallholder farms (Table 2). This indicates that GBS had better opportunities to exploit G x E than CBS. This could be explained by the fact that in GBS, there is either less or no emphasis on own performance under ideal nucleus environments and high selection accuracy in a harsh environment for both sexes [51]. The GBS possesses the ability to move the emphasis to performance in field conditions [51], confirming the use of GBS to select chicken in the nucleus for the improvement of their offspring performance under the field conditions [14, 21]. This shows that GBS is much better in increasing resilience and reducing the environmental sensitivity of animals [51]. Being a good predictor of performance of IC in smallholder farms, GBS would thereby be a beneficial breeding scheme to adopt in Rwanda. This is because so far there is no information about G X E and the majority of IC farmers are smallholder rearing their chickens under extensive management [45, 42].
The current study revealed that the rate of inbreeding was lower in GBS than CBS (Table 2). The reason for this lower rate of inbreeding is that, by offering information on Mendelian sampling terms, genomic markers distinguish relatives, comprising full sibs, which decreases the likelihood of co-selection of relatives [12, 15]. Marker information is entirely heritable and has no residual variance since it is supposed that genotypes can be observed without error [65], thus, as heritability increases, accuracy of selection increases thereby rate of inbreeding decreases [15]. This infers that GBS should be adopted by breeding programs at the earliest possible time to improve genetic gain over the shortest period and to restrain the inbreeding rate as per the recommendation by [60, 18]. The challenge of implementing GBS, however, was the exorbitant cost of genotyping [68]. Fortunately, the cost of genotyping is, however, becoming low [67] and it is therefore supposed that genomic selection would be largely applied in poultry breeding programs even in Rwanda.
Genetic gain realised for individual traits differed between CBS and GBS (Table 3), and also between nucleus and commercial. Genetic gain per generation for BW observed in this study was positive and higher than 69.92g of IC in Tanzania [37], 38.72g of IC from Kenya [60] and 72.00g of Ethiopian chicken [81]. Genetic gain for EN obtained in this study was lower than 1.36 IC eggs from Kenya [60], 2.45 eggs for Tanzanian medium IC [37], 2.6 eggs for the Indian chicken ecotype [75] and 3.10 eggs for chicken in Nigeria [59]. These dissimilarities could be attributed to either population size considered, selection procedure used, or selection intensities employed [71]. This is because the most important factors which affect genetic gain include effective population size, accuracy of selection, and selection intensity [59]. Optimal genetic gain could be realised by maximising these aforementioned factors [9]. Regrettably, with restricted resources, all these factors cannot be maximised concurrently [59]. For instance, increasing the intensity of selection lowers an effective population size and leads to a reduced genetic gain [50]. Correlated genetic gain achieved for EW was lower in the nucleus than in commercial in the two breeding schemes, contrarily to BW and EN. This difference could be caused by the negative phenotypic and genetic correlation between EW and EN [57]. This suggests that selection of IC for EN would compromise egg size. However, eggs from IC can be marketed with their present low weight with no problem [37]. Furthermore, egg consumers are currently not ready to pay much more for bigger eggs [4, 55]. The negative genetic gain for AbR realised in the two breeding schemes was a sign of decreased AbR combating Newcastle disease and thereby low immunity. Observed decline in correlated response for AbR in the two breeding schemes suggests that intense selection for BW and EN would compromise the immune system of fowls. In poultry as in other livestock species, the antagonistic relationship between functional and production traits has been demonstrated [32, 43, 31]. The current study suggests that selecting IC for BW together with EN and selection of IC for EW alongside AbR and crossing them would, therefore, assist in the utilisation of hybrid vigour and complementarities.
4.2. Optimum Nucleus Size
Increase in nucleus size resulted in an upturn in genetic gain and a decrease in the rate of inbreeding in both CBS and GBS (Fig. 2). The cause for this trend was that the increase in nucleus size would increase the number of selected unrelated parents and hence reduce the rate of inbreeding [30]. In GBS, by increasing the amount of genomic information and the size of the reference population improves the genetic gains [18]. Rates of inbreeding may be reduced by increasing the flock size and the numbers of sires and dams selected [78]. Increasing flock size, however, could increase the cost of a nucleus flock program [33]. Researches have, therefore, recommended to focus on ways to obtain a maximum genetic response in nucleus herds/flocks of small size and to limit inbreeding at the same time [7]. Through this study, there was, therefore, a sharp increase of genetic gain up to a nucleus size of 960 and 303 hens with sex ratios of 1:5 and 1:100, respectively and then, a diminishing increase as the nucleus size gets much larger. Since this study also revealed an effect of sex ratio on nucleus size, the optimal nucleus size in Rwanda, therefore, could be considered as one ranging between 303 and 960 hens with sex ratios varying between 1:100 and 1:5, respectively.
In addition, the current study showed that the increase in mating ratio resulted in a great genetic response in both CBS and GBS (Fig. 3). This trend may be because, by increasing the mating ratio, the selection intensity of males increased, leading to the improvement of response to selection due to a positive correlation between selection intensity and genetic response [8]. This result could justify the use of semen from one cock to inseminate one hundred (100) hens using artificial insemination reproductive technology [48]. This later approach allowed the quick dissemination of genetic material from a small number of superior males to a high population of females [48]. Besides, a higher mating ratio yielded a greater rate of inbreeding in CBS. This could be because, once the number of parents is small, the rate of inbreeding is expected to be higher [73]. Inbreeding leads to increased homozygosity within the population, resulting in reduced genetic variance [83]. Increase in inbreeding is of great burden in livestock breeding programs because it decreases the long-term response to selection due to reduced genetic variance in a population [14]. Even if inbreeding cannot be avoided [22], the mating ratio would, therefore, depend on an optimal rate of inbreeding not more than 1% per year recommended by FAO [69] to void the loss of merit and genetic variance. This study revealed that an increase in the mating ratio led to a reduction in the rate of inbreeding in GBS. This is because using genomic information reduces the probability of co-selection of relatives in the parents [12, 15]. As recommended by [47], the nucleus size and mating ratio that capitalize on genetic gain while limiting the rate of inbreeding in IC would further result in an optimal genetic contribution of the current generation to next generations.