Genetic characterization of Indonesian sorghum landraces ( Sorghum bicolor (L.) Moench) for yield traits

: Sorghum ( Sorghum bicolor (L.) Moench) is the fifth most produced cereal crop in the world. The use of sorghum is very diverse and most parts of the plant, including stem, leaves, grain, panicles, stem juice, and bagasse, can be utilized as human food, animal feed


Introduction
Sorghum (Sorghum bicolor (L.) Moench) is a member of the Poaceae family.The cultivation of sorghum began in ancient Africa around 8000 BC, with the East Sudanian savanna as the center of origin, then spread and became an important food crop in China and India [1].At present, sorghum is cultivated throughout the world from nutrient-low soils to fertile soils in tropical to temperate areas [2].Currently, sorghum is the fifth most-produced cereal crop in the world, after corn, rice, wheat, and barley [3].The area of sorghum cultivation reached 41.75 million ha worldwide, with the largest areas in Sudan (7.5 million ha) and Nigeria (5.7 million ha).The global sorghum production as of 2023 was 59.92 million t with an average productivity of 1.44 t ha -1 [4].According to the report by the United States Department of Agriculture (USDA), the largest sorghum-producing country is the United States (8.18 million t, followed by Nigeria (6.70 million t), Sudan (5.00 million t), Mexico (4.80 million t) and India (4.40 million t) [4].
Sorghum can be utilized for various uses.Most parts of the sorghum plant, including stem, leaves, grain, panicles, stem juice, and bagasse, are utilized as human food, animal feed, and material/source for industry and bioenergy production [5][6][7].Sorghum grains as food have good nutritional value [8,9].As livestock feed, sorghum produces grains that can be a nutritionally equivalent substitute for corn, as well as dried leaves and stems as good sources of dietary fiber [10].As industrial raw materials, a certain type of sorghum (sweet sorghum) has been developed for the manufacturing of liquid sugar and syrup [11], beer [12], and ethanol [7,13,14].The ethanol produced from sweet sorghum can be alternative to fossil fuels [7,14,15], and the bagasse can be utilized for the production of particle boards and bio-pellets [7,16,17].
Sorghum cropping has the potential to support food and energy programs in Indonesia [9].The area of sorghum fields in Indonesia was only 26,306 ha in 2012-2013 [18].The major sorghumproducing areas consist of nine provinces, including East Nusa Tenggara (58.3% of cultivation area), Southeast Sulawesi (15.2%),South Sulawesi (12.9%),East Java (8.4%), and several others (< 4%).However, the productivity was not very high at that time, in the range of 1-2 t ha -1 [18].There is very limited update data on sorghum production and productivity in Indonesia since sorghum is not the priority crop in Indonesia.According to the Press Release of the Coordinating Ministry for Economic Affairs of Indonesia, number HM.4.6/412/SET.M. EKON.3/08/2022, the area of sorghum field in Indonesia in 2022 was estimated at 4,335 ha and the sorghum production in six provinces was 15,243 t.The estimation of sorghum productivity in Indonesia was 3.36 t ha -1 .To provide enough material for the downstream industries, improving productivity is needed.In addition, sorghum production should be expanded to the marginal lands that are not suitable for the cultivation of ordinary crops in the central and eastern parts of Indonesia to avoid land-use competition between crop commodities [9].Efficient production of sorghum in such marginal lands requires the selection of accessions with pest and disease resistance or abiotic stress tolerance.
The main repository for world sorghum germplasm is the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) which is headquartered in India.A total of 39,234 accessions from 93 countries have been recorded by the Institute [19].In Indonesia, the Agricultural Genebank of the Ministry of Agriculture conserves 259 sorghum accessions collected from local regions in Indonesia or contributed from abroad gene banks and researchers [20].The collected germplasms are diverse and potentially useful as the source of traits to improve grain yield for food production, biomass yield and quality for fodder production, or stem sugar content, lignin, and cellulose content for industrial application [21][22][23][24][25][26].In this study, we characterized 40 Indonesian local sorghum accessions based on morphological and productive point of view [27], to identify suitable and promising accessions for further utilization.The characteristics analyzed include plant size, biomass production, seed/grain production, and sugar content.These parameters are critical for the performance of sorghum as food, feed, bioenergy feedstock, and other industrial materials.The results of this study will be useful for the government, researchers, farmers, a n d business/industry sectors in Indonesia and many other countries.

Time and location of the cultivation
Cultivation of 40 Indonesian local sorghum accessions was conducted from April to August 2017 at Cibinong Science Center and Botanic Gardens of the Indonesian Institute of Science (LIPI), which has been changed into National Research and Innovation Agency (BRIN).The area is in Bogor Regency of West Java Province of Indonesia, with the altitude of 250 m above sea level.The monthly average minimum and maximum temperature during the research were 22.96 and 31.95℃, respectively; the monthly average of air humidity ranged between 76.61% and 85.29%, a monthly average of wind speed was from 1.43 and 5.35 km h -1 ; and the monthly average rainfall ranged between 179.2 and 401.9 mm [28].The soil type of the study location was alfisol with the chemical properties shown in Appendix 1.

Plant cultivation
We obtained the seeds of 37 Indonesian local sorghum accessions and the seeds of 3 (three) standard sorghum accessions from the Agricultural Genebank of the Ministry of Agriculture in Bogor-West Java, and the sub-genebank in Maros-South Sulawesi, respectively.A complete list of the 40 sorghum accessions used in this study is presented in Appendix 2. We applied a Completely Randomized Design for the study with one plot for each accession of sorghum tested.Plant cultivation was conducted in test plots sized 5 m × 5 m and placed 4 m apart.Each accession of sorghum was cultivated in one plot from a total of forty test plots.Two seeds were sown per hole, with spacing 25 cm in the rows that were 75 cm apart in the plots.The fields were plowed and applied with 10 kg of compost before planting.Urea (150 kg ha -1 ), triple superphosphate (150 kg ha -1 ), and potassium chloride (150 kg ha -1 ) were applied as basal fertilizers at the time of planting, and urea (150 kg ha -1 ) were applied as top dressing at 1 month after planting.

Measurement of parameters
Plant growth parameters and production-related parameters were measured.The plant growth parameters were analyzed during the flowering period, including plant height (PH), leaf number (LN), leaf length (LL), leaf width (LW), panicle length (PL), and panicle stalk length (PSL).The productionrelated parameters were analyzed at harvest, including 100 grain or seed weight (100SW), total grain weight (TSW), panicle weight (PW), fresh and dry weight of plant biomass (FWB and DWB), and sugar content in stem juice (SC).Plant growth parameters, production-related parameters, and sugar content were measured on three randomly selected individuals from each plot.Sugar contents were measured in triplicate by taking juice from the middle part of the stem using the digital refractometer (Minolta, Palette Series, ATAGO Limited Company).

Statistical analysis of the data
Obtained data were subjected to descriptive statistics, analysis of variance (ANOVA) with Posthoc Tukey, and multivariate analysis using Minitab ® 19 Software (Minitab Inc., State College, Pennsylvania, USA), including correlation, principal components (PC), and cluster analysis.The PCA-biplot analysis was performed using the Biplot-Excel Program.In this study, the 40 Indonesian local sorghum accessions were characterized based on multiple growth and production-related parameters.Table 1 shows the degree of association among the parameters estimated by Pearson correlation coefficients, with significant correlation shown in bold letters.In general, plant biomass (FWB and DWB) showed a significant positive correlation between the size-related parameters, including PH, LN, and LL.Positive correlations were also found among grain production-related parameters, including 100SW, TSW, and PW.No significant correlation was found between SC or PSL and other parameters (Table 1).A positive correlation between PH and biomass production (either dried or fresh) has also been reported in previous studies [26,[29][30][31][32].These and the present results together suggest that taller sorghum has a higher capability of biomass production.The relationship may help to identify the genotype suitable for biomass utilization.On the other hand, no clear correlation was found between SC and the other parameters examined in this study, thus additional parameters need to be explored as aids in the selection of sorghum lines suitable for sugar production.

Cluster analysis
A cluster analysis using all measured data separated the 40 accessions into three clusters (Figure 1).The accessions belonging to each cluster are listed in Table 2. Local sorghum accessions from the same origin tended to be placed in the same cluster.Eight accessions from Belu of Nusa Tenggara Timur Province (Butter Ainarup 1 and 2, Butter Bebelit 2, Butter Biara, Butter Krek, Butter Mean, and Butter Nean Reket A and B), and three accessions from Nusa Tenggara Barat province (Lokal Bima 1, 2, and 3) were grouped in cluster 1 (Figure 1).Two out of three accessions of Selayer (Selayer 1 and 3) also belonged to cluster 1 (Figure 1).Cluster 2 mainly consisted of the accessions from Central Java, including five out of six accessions from Demak (Demak 1, 2, 3, 4, and 5), Kempul Putih 64 K6, and Kempul Putih 82 R10 (Figure 1).Moreover, almost half of all accessions in cluster 3 were collected from West Java including Coley, Keler, Kolot, and RGV (Figure 1).According to the mean values of the parameters within each group, the accessions in cluster 1 were characterized by lower LW, TSW, and PW compared to those of the members in clusters 2 and 3 (Table 3).The results indicated that the accessions in cluster 1 had smaller sizes of leaves and lower grain production.The accessions in cluster 2 were characterized by higher TSW and PW compared to those of the members in clusters 1 and 3 (Table 3).These results indicated that the accessions in cluster 2 had a higher capacity for grain production.The accessions in cluster 3 were characterized by higher PH, FWB, and DWB compared to those of the members in clusters 1 and 2 (Table 3).The results indicated that the accessions in cluster 3 had a higher capacity for biomass production.
The improved cultivars included in this study (Super 1, Super 2, and Kawali) were separated into different clusters.The cultivars Super 1 and Super 2 were originally developed as those having higher sugar content [33].However, the observed variation in SC among the 40 accessions were rather small in this study, thus SC did not become the factor discriminating the cultivars in this study.Rather, they appeared to be separated according to the capacity of biomass production.The cultivars Super 1 and Super 2 produced higher amounts of biomass (FWB and DWB), whose DWB ranked 4 th and 7 th among the 40 cultivars, respectively.

PCA-Biplot analysis
The grouping of 40 sorghum accession was further examined by PCA-biplot (Figure 2).The first and second principal component explained 32.4% and 19.6% of the variation in the traits, respectively.The accessions belonging to the same group in the clustering analysis were placed in proximity on the scatterplot.The result supports the validity of the grouping by clustering analysis.The vectors with positive value along the first axis include FWB, DWB, LL, LN, PH, and SC.The vectors with positive value along the second axis include TSW, PW, 100SW, and LW.The vectors with negative value along the second axis include PL and PSL (Figure 2).Thus, the first and the second axis are likely to indicate the capacity for biomass production and grain production, respectively.Hence the cluster 3 can be characterized as the group with high seed production, whereas the cluster 2 is characterized as the group with high seed production but not high biomass production.The members in cluster 1 can be regarded as those with neither high seed production nor high biomass production.The result is in accordance with the characteristics predicted in the clustering analysis (Table 3).Separation of the clusters 2 and 3 on the first axis suggests that most accessions with high potential for grain production are not high in biomass production, as the result of a trade-off between the two traits.The accession 34 (Rio) was placed in cluster 1 by cluster analysis, whereas it was placed closer to those in cluster 2 by PCA-Biplot (Figure 2, Table 2).This accession showed lowest biomass productivity but higher grain productivity among the 40 accessions tested, with DWB and TSW ranked 40 th and 4 th , respectively.These characteristics may have caused the fluctuation in classification.Note: Each of the dots represents an individual accession, with blue, magenta, and green representing the accession belonging to clusters 1, 2, and 3, respectively.

. 4 . Analysis of variance
The results of the analysis of variances of all parameters measured among 40 sorghum accessions were significantly different (Appendixes 3 and 4).Significant higher values of the parameter related to biomass production (FWB and DWB) were found for the accessions Coley, Keler, Lao, Lokal Kaltim, and Super 1 as compared to those of the other accessions (Appendix 4).On the other hand, significantly higher values of the parameter related to grain production (100SW, TSW, PW) were found for the accessions of Demak 1, Demak 2, Demak 3, Demak 4, Demak 5, Kempul Putih 82 R10, Lao, Lokal Kaltim, Nean Reket, and Rio as compared to those of the other accessions (Appendix 4).The accession with a significantly higher value of sugar content (SC) parameters as compared to that of most of the other accessions was Rio accessions (Appendix 4).

Potential usefulness of local sorghum accessions
The results of the characterization of 40 Indonesian local sorghum accessions in this study could be utilized for many purposes, including the conservation and maintenance of genetic resources for breeding or future usage.So far, many sorghum germplasm have been collected by the Agricultural Genebank, including cultivars, landraces, and inbred lines with various phenotypes and origins [34].Sorghum accession with high potential for biomass production is promising as livestock fodder or the material for bioenergy production [7].The results of this study suggest that the accession with high PH, LN, LL, FWB, and DWB could have a high potential for biomass production (Table 1, Appendix 4).The accessions having such characteristics are Coley, Keler, Lao, Lokal Kaltim, and Super 1 (Figures 1 and 2, Table 2, Appendix 4), with the value of DWB ranging 20.69 t ha -1 -25.28 t ha -1 .These accessions were placed into cluster 3 based on cluster and PCA-biplot analyses (Figures 1 and 2).
Potential for grain production is also an important trait in the development of new varieties.Sorghum grains can be utilized not only as food or feed but also as a source of starch for ethanol production [7].The top ten accessions with high potential for seed production as judged by TSW included Demak 1, Demak 2, Demak 3, Demak 4, Demak 5, Kempul Putih 82 R10, Lao, Lokal Kaltim, Nean Reket, and Rio, with values were 3.72 t ha -1 -5.02 t ha -1 .The values are comparable with those of improved sorghum varieties released between 1960 and 2001 in Indonesia, including No 46, No 6C, UPCA-S2, UPCA-S1, KD4, Hegari Genjah, Mandau, and Numbu [18].Most of the high grain producers identified in this study belonged to cluster 2, whereas Lokal Kaltim (ranked 2 nd ) and Lao (ranked 5 th ) belonged to cluster 3, which is characterized as high biomass producers (Figures 1 and 2, Table 2, Appendix 4).The DWB of Lokal Kaltim and Lao indeed ranked second and first among the 40 accessions, respectively (Appendix 4).As discussed above, there appears to be a trade-off between the production of biomass and grain in general (Figure 2), whereas these accessions exhibited high potential in both aspects.

Conclusions
The 40 local sorghum accessions could be classified into 3 groups based on their biomass and seed productivity.The clustering tends to reflect the provenance of the accessions.Five accessions belonging to cluster 3 had high biomass productivity and appeared to be promising for use as feedstock for biomass energy production and forage, or as parent lines for further improvement by breeding.The accessions include Coley, Keler, Lao, Lokal Kaltim, and Super 1.On the other hand, 10 accessions in cluster 2 could be utilized for seed production application, i.e.Demak 1, Demak 2, Demak 3, Demak 4, Demak 5, Kempul Putih 82 R10, Lao, Lokal Kaltim, Nean Reket, and Rio.Lokal Kaltim and Lao are particularly promising in that they combine high biomass yield with grain yield.The identification of genes associated with these agronomically useful traits using these lines will help in the selection of superior genotypes.Further studies on the characteristics of plant resistance to biotic and abiotic stresses of the plants are also required.

Use of AI tools declaration
The authors declare that they have not used Artificial Intelligence (AI) tools in the creation of this article.Volume Note: PH = Plant Height, LN = Leaf Number, LL = Leaf Length, LW = Leaf Width, PL = Panicle Length, PSL = Panicle Stalk Length, 100SW = Weight of 100 Grains, TSW = Total Grain Weight, PW = Panicle Weight, FWB = Fresh Weight of Plant Biomass, DWB = Dry Weight of Plant Biomass, SC = Sugar Content.* and ** significant at p ≤ 0.05 and p ≤ 0.01, respectively.

Figure 1 .
Figure 1.Hierarchical clustering analysis of 40 Indonesian local sorghum accessions.Note: The Dendrogram was constructed by the Ward method of cluster analysis based on Euclidean distances.
59 a Note: Shown are the mean ± standard deviation of values for the accessions in each cluster.The different letter indicates significant difference between the clusters (p < 0.05, Tukey HSD test).PH = Plant Height, LN = Leaf Number, LL = Leaf Length, LW = Leaf Width, PL = Panicle Length, PSL = Panicle Stalk Length, 100SW = Weight of 100 Grains, TSW = Total Grain Weight, PW = Panicle Weight, FWB = Fresh Weight of Plant Biomass, DWB = Dry Weight of Plant Biomass, SC = Sugar Content.

Table 1 .
Correlation of the vegetative and production variables in 40 Indonesian local sorghum accessions.

Table 2 .
Member of three groups of 40 Indonesian local sorghum accessions when mapping the biplot analysis with the cluster analysis.

Table 3 .
Means of sorghum vegetative and production variables in each cluster.

Table 14 .
Sugar Content of Stem Juice (SC) Parameter.

Table 17 .
The parameters of panicle weight (PW), fresh weight of plant biomass (FWB), dry weight of plant biomass (100SW) and sugar content of stem juice (SC).