Genetic Parameters of Agronomic Traits in Sweetpotato Accessions

Germplasm as a source of genes in sweetpotato breeding requires information on appearance and genetic parameters. The objectives of this research were to determine the performance and genetic parameters of sweet potato accessions. The research was conducted at Kendalpayak Research Station, Malang, East Java, Indonesia. The materials used were thirty sweet potato accessions from Indonesian Legumes and Tuber Crops Research Institute (ILETRI) germplasm collection. The research was arranged in a Randomized Block Design (RBD) with two replications. The variables observed included: the vines length, the weight of vines, the number and weight of the saleable root per plot, the number and weight of the non-saleable root per plot, the number and weight of root per plant, the root yield, the harvest index, and the dry matter content. The results of ANOVA showed a significant difference among the tested genotypes in almost all traits observed except on weight of non-saleable root. PCV estimation was higher than GCV estimation for all the observed characters. The weight of the saleable root per plot, the weight of root per plant, and the root yield that showed a wide range of PCV and GCV as well as high broad-sense heritability indicated that these traits had additive gene effect and more reliable for effective selection. The broad GCV in a population is effective for selection to obtain the superior variety.


INTRODUCTION
Sweet potato (Ipomoea batatas L.) is one of the important staples food in Indonesia. It is the fourth source of carbohydrate after rice, corn and cassava. It also contain fiber, vitamins, minerals, antioxidants, and have a low glycemic content. Sweet potatoes play an important role in the supply of industrial raw materials and animal feed (Wera et al., 2014;Pradhan et al., 2015). The development and improvement of sweet potato's productivity is needed to meet those needs. Productivity improvements can be made through the breeding programs.
Breeding programs will succeed if supported by information of economic value, wide diversity, and high inheritance of the character to be corrected. Therefore, understanding the diversity of sweet potato genotypes based on agronomic traits is very important in planning sweet potato breeding programs and determining effective selection criteria (Ngailo et al., 2016;Selaocoe et al., 2019). Selection in breeding programs is the basis of all improvements to get new superior varieties. Efficient selection will be obtained by using several genetic parameters and heritability as considerations. According to Palumbo et al., (2019) and Irwan et al., (2019), wide genetic diversity is one of the conditions for an effective selection program, and selection for a desired character will be more meaningful if the character is easily inherited.
Another important component that determines the success of a variety assembly program is information about inheritance (heritability values) and characters that are positively correlated with root yields. Breeders need this information to determine their selection strategies and criteria. Studies on parameters genetic and heritability has been done by researchers. Shaumi et al., (2012); Madawal et al., (2015); Rahajeng and Rahayuningsih (2016); EL-Sharkawy (2019) reported that jumlah dan bobot umbi serta hasil umbi adalah the characters with the high and moderate heritability and genetic advance can be considered for direct selection for sweet potato improvement.
The objectives of this research were to determine the performance and genetic parameters of sweet potato accessions from the germplasm bank of Indonesian Legumes and Tuber Crops Research Institute (ILETRI). Information about the performance and genetic parameters of sweet potato accessions from this study is expected to be utilized for sweet potato breeding programs.

METHODS
The research was conducted in February-June 2017 at Kendalpayak Research Station, Malang, East Java, Indonesia. Kendalpayak Research Station lies at 8° 2′ 56.4″LS 112° 37′ 30″BT with an altitude of 445 m a.s.l. The average annual rainfall was 2191 mm with a minimum/maximum mean air temperature of 17.5/30 o C. The soil was classified as Entisol and the textural class was clay with pH of 5.8.
The material used was thirty sweet potato accessions from Indonesian Legumes and Tuber Crops Research Institute (ILETRI) germplasm collection. The research was arranged in a Randomized Block Design (RBD) with two replications. Each accession was planted on 5 m x 1 m (single row), which accommodated 20 plants (the spacing between plants was 25 cm). Fertilizer (300kg/ha of NPK Phonska) was applied at planting (2/3 dose) and 5 weeks after planting (1/3 dose). Weeding was done at 4, 7, and 10 weeks after planting. Irrigation, pest, and disease control was applied as needed. Harvesting was done after 4 months of planting. The variables observed included: the vines length (cm), the weight of vines (kg/plot), the number of the saleable root per plot, the number of the non-saleable root per plot, the number of root per plant, the weight of the saleable root per plot (kg), the weight of the non-saleable root per plot (kg), the weight of root per plant (kg), the root yield (t/ha), the harvest index, and the dry matter content (%).
Data were analyzed by analysis of variance (ANOVA) using PKBT-STAT 1.0 program. Genetic parameter analysis (genotypic and phenotypic coefficients of variation, heritability, and genetic advances) was performed according to Syukur et al. (2009) and Demelie and Aragaw (2016). The deviation of genetic variance was used to determine the criteria of genetic variability.

Analysis of Variance (ANOVA) and Mean Performance of Genotypes
The ANOVA test for eleven characters showed significant differences among the genotypes for almost all characters observed except weight of the non-saleable root (Table 2). This may indicate that each accession showed a different genetic, especially for the vines length, weight of vines, number of the saleable root per plot, the number of root per plant, the weight of the saleable root per plot, the weight of the non-saleable root per plot, the weight of root per plant, the root yield, the harvest index, and the dry matter content and also had a wide variability among genotypes.
The performance of the agronomic traits in thirty sweet potato accessions are presented in Table 3. Each trait of the accessions had a wide range of mean values. MLGI 0037 produce the highest yield, harvest index, and weight of root while MLGI 0031 showed the lowest value. The root dry matter content of thirty accessions had range between 23.64% (MLGI 0014) and 34.93% (MLGI 0006) with an average 28.70%. Dry matter content of root is a critical parameter in the selection because it can be used as an indicator of root quality. According to Kathabwalika et al., (2013), the root dry matter content indicates mealiness in the roasted or boiled sweet potato, so that it becomes an important quality parameter in the food processing industry and determines consumer preferences. The farmer preference of the root dry matter content is > 25% (Mbah and Eke-Okoro, 2015). Meanwhile for industry, the root dry matter content preference is > 30% (Rukundo et al., 2013).  0.14 ** 17.00 ** 0.01 0.33 Note: ** significant at p< 0.01, * significant at p< 0.05, ns= non significant

Estimation of Genetic Variance Component
In breeding programs, selection is the main activity to obtain superior varieties. Selection will run effectively if a population with a broad genetic diversity is available. So the opportunity to get the desired traits increases. Table 4 shows eight characters (weight of vines, number and weight of root perplant, weight of saleable and non-saleable root perplot, root yield, harvest index, and dry matter content) of eleven characters observed which had broad genotypic coefficient of variation (GCV). These results are similar to the results of a study by Badu et al., (2017) and Sharavati et al., (2018) which obtained broad GCV for weight of vines, weight of root, and root yield. Vine length and number and weight of nonsaleable root per plot showed narrow GCV. In breeding program, broad GCV will expand the opportunity to improve these characters through selection, because it indicates the large amount of variation (Badu et al., 2017;Narasimhamurthy et al., 2018). A rigorous selection method should be done to select the characters with narrow GCV (Addisu et al., 2013;Kuswantoro et al., 2018).
Phenotypic coefficients of variation values were higher than genotypic coefficient of variation values with slight difference values for all the observed characters. Akinwale et al., (2010) and Baafi et al., (2016) stated that the diversity is also influenced by environmental factors besides the genetic factors. The slight difference in value between PCV and GCV shows that the influence of genetic factors is more dominant than the influence of environmental factors. The research of Demelie and Aragaw (2016) and Rahajeng and Indiati (2018) also showed the same results, PCV values were greater than GCV with a slight difference in values. If the value of PCV and GCV has a significant difference, it means that environmental factors have a high influence.
In addition to the coefficient of variance, information about inheritance is also important to determine the criteria for efficient selection. Kuswantoro et al., (2018) stated that the coefficient of variance only shows the variability of genotypes of the observed characters but does not provide information about the proportion of inheritance. Therefore, the value of heritability needs to be known to determine the pattern of inheritance. Heritability indicates the amount of influence of genetic factors or environmental factors on a character. The high heritability value shows that the character is more influenced by genetic factors. A character that has a high heritability value can be used as an effective selection criteria in the early generations (Chahal & Gosal 2010;Afuape et al., 2015;Dewi et al., 2019).     In this study, heritability value showed that almost all of the characters observed have high broad-sense heritability except for the weight of the non-saleable root (Table 5). This result means that the phenotypic appearance of the other 10 characters are more influenced by genetic factors rather than by environmental factors. While the medium heritability value on the character shows that the influence of the environment and genotype is at the same level. The similar result were obtain by Dewi et al., (2019) that reported the weight of vines, number of large root, and weight of root that showed high heritability values. Root yield, harvest index, and root dry matter content also show high heritability values on study by Shumbusha et al., (2019) High heritability coupled with broad GCV indicated that the characters had additive gene effect and more reliable for effective selection. In this study, the weight of the saleable root per plot, the weight of root per plant, and the root yield showed broad GCV and high broad sense heritability. These results are in agreement with study by Wera et al., (2014), Rahajeng and Rahayuningsih (2016), and Narasimhamurthy et al., (2018).
The broad GCV in a population is effective for selection to obtain a superior variety. PCV estimation was higher than GCV estimation for all the observed characters. Inheritance information is important to determine the criteria for efficient selection in addition to the coefficient of variance. Almost all of the characters observed have high broad-sense heritability except for the weight of the non-saleable root.
Results of this study showed that based on broad GCV dan high heritability, the weight of the saleable root per plot, the weight of root per plant, and the root yield are more reliable for effective selection. The benefit of this study are the 30 accessions can be utilized in sweet potato breeding programs especially for characters which have a broad GCV (the weight of the saleable root per plot, the weight of root per plant, and the root yield) since they can be combined as crossing parents and used these characters as the selection criteria.

CONCLUSIONS
The results of ANOVA showed the significant difference among the tested genotypes in almost all traits observed except on weight of non-saleable root. PCV estimation was higher than GCV estimation for all the observed characters. The weight of the saleable root per plot, the weight of root per plant, and the root yield showed a wide range of PCV and GCV as well as high broad-sense heritability that indicated these traits to have additive gene effect and more reliable for effective selection. The broad GCV in a population is effective for selection to obtain superior variety. A total of 30 accessions can be utilized in assembling varieties especially for characters which have a broad GCV value.