Genetic relationships in Saccharum complex germplasm collections based on morphological and molecular markers

Sugarcane has a high degree of heterozygosity and is a cross-pollinator, so information about the genetic relationship between the accessions in germplasm collections is very important for selecting the prospective parent in crossbreeding. This research aims to determine the phylogenetic relationship of 24 Saccharum complex accessions and to verify the grouping of accessions using 37 morphological and three microsatellite molecular markers. Interpretation of morphological and molecular data was obtained from the analysis using the NTYSYpc-2.02i program. The results show that within the 24 accessions analyzed using morphological markers, some accessions did not cluster as the classification at the conservation time. This difference is due to the morphological markers, so the value of genetic similarity among accessions analyzed is increased. In contrast, the grouping of molecular markers shows that each accession was grouped according to the classification at the conservation time. These accessions had a low genetic similarity of 0.20 with a broad genetic distance of 0.80. This broad genetic distance indicates that the twenty-four accessions have a distant genetic relationship with one another, so that the genetic diversity of these accessions is relatively high. The high genetic diversity in germplasm collections improves its potential as a crossing parent to obtain a high heterosis effect. [


Abstract
Sugarcane has a high degree of heterozygosity and is a cross-pollinator, so information about the genetic relationship between the accessions in germplasm collections is very important for selecting the prospective parent in crossbreeding. This research aims to determine the phylogenetic relationship of 24 Saccharum complex accessions and to verify the grouping of accessions using 37 morphological and three microsatellite molecular markers. Interpretation of morphological and molecular data was obtained from the analysis using the NTYSYpc-2.02i program. The results show that within the 24 accessions analyzed using morphological markers, some accessions did not cluster as the classification at the conservation time. This difference is due to the morphological markers, so the value of genetic similarity among accessions analyzed is increased. In contrast, the grouping of molecular markers shows that each accession was grouped according to the classification at the conservation time. These accessions had a low genetic similarity of 0.20 with a broad genetic distance of 0.80. This broad genetic distance indicates that the twenty-four accessions have a distant genetic relationship with one another, so that the genetic diversity of these accessions is relatively high. The high genetic diversity in germplasm collections improves its potential as a crossing parent to obtain a high heterosis effect.

Introduction
Sugarcane is one of the most important industrial crops in tropical and subtropical regions because of its ability to store high carbohydrates to produce sugar and biofuel. Sugarcane is an economically important agricultural crop cultivated in over 90 countries, contributing to nearly 80 % of global white sugar production (Raj et al., 2016;Diederichs et al., 2016;Aitken et al., 2018;Pocovi et al., 2020).
Currently, the national sugar production has not been able to meet consumption needs due to several constraints, including low sugarcane and sugar productivity. One of the efforts to overcome this food problem is to intensify breeding activities to obtain superior sugarcane varieties. The assembly of new superior sugarcane varieties is highly dependent on the availability of genetic diversity of Saccharum spp.
The genetic diversity of Saccharum spp. is abundant in Indonesia because this country is known as one of the original centers of Saccharum (FAO, 2019). The collection of sugarcane germplasm in the Indonesian Sugar Research Institute (ISRI) began in the 1890s by Dutch researchers through obtaining accessions of sugarcane from within the Indonesian archipelago and introducing clones from abroad (ISRI, 1997). Lamadji (1994) reported that six expeditions conducted in Indonesia to explore wild sugarcane in Kalimantan, Sulawesi, Maluku, North Sumatera, and Irian Jaya (West Papua) in 1976, 1984, 1985, and 1995 In 2014, the volume of sugarcane germplasm collections managed and conserved by ISRI consisted of 5,294 accessions. The collection consists of a group of native cane (Saccharum officinarum), wild cane (Saccharum spontaneum, Saccharum robustum, Saccharum barberi, Saccharum sinensis, and Saccharum edule), wild relatives of Saccharum (Erianthus and Miscanthus), and Saccharum hybrid (Mirzawan et al., 2014). The modern sugarcane cultivars originated from the Saccharum complex, which consists of two wild Saccharum species (S. spontaneum, and S. robustusm), four cultivated species (S. officinarum, S. sinense, S. barberi, and S. edule), and four interrelated genera (Erianthus, Miscanthus, Narenga, and Sclerostachya) (Todd et al., 2014).
Sugarcane is a polyploid plant and has a very complex genome. The Saccharum species present large genomes and variations in the number of chromosomes (Liu et al., 2016;Vieira et al., 2018). The complexity of the Saccharum genome is inherited by modern sugarcane cultivars, in which its levels of ploidy, often aneuploidy, and have a large genome size of about 10 GB (Vieira et al., 2018;Zhang et al., 2018).
Pollination of sugarcane is both openpollinated and cross-pollinated. Individual plants have a high degree of heterozygosity and the appearance of superior varieties caused by heterosis (Birchler & Riddle, 2014). With these types of pollinations, sugarcane will produce genetically distinct offspring: between individual offspring, between offspring and their parents, between generations of one offspring and other offspring generations. To maximize the utilization of the sugarcane germplasm collection, it is necessary to evaluate and analyze genetic relationships between the accessions. Genetic relationships between collected germplasm can be evaluated and analyzed through morphological and molecular markers. This will facilitate the selection process of crossing parents and optimize selection procedures (Khan et al., 2009).
An important step before crossing is the selection of the correct parent cross to obtain superior offspring. The strategy for selecting parents can be carried out by characterizing the germplasm with morphological and agronomical markers. However, errors can occur because the vegetative traits are influenced by the environment and often do not reflect the genetic diversity of actual Saccharum accessions (Manechini et al., 2018;Medeiros et al., 2020). Therefore, molecular markers can be used as a complementary tool to characterize morphologically and to identify accessions accurately and reliably so that the combination of accessions as cross parents is suitable for breeding purposes (Marconi et al., 2011;dos Santos et al., 2012;Singh et al., 2018).
This study aims to determine the genetic relationship of 24 clones of Saccharum complex members based on morphological and microsatellite molecular markers and to verify the grouping of 24 Saccharum complex clones based on microsatellite markers.

Materials and Methods
This research was conducted in the ISRI's germplasm collection station and Biotechnology Laboratory in the 2017/2018 planting season. Twenty-four accessions of ISRI's sugarcane germplasm were members of the Saccharum complex, consisting of S. officinarum, S. barberi, S. sinense, S. robustum, S. spontaneum, Erianthus, Miscanthus, and Saccharum hybrids as an outgroup used in this study (Table 1).

Morphological characterization
Morphological characterization was carried out directly in the ISRI's germplasm collection station when the plants were ± seven months old. Each accession was planted in 2 rows, and each row consisted of 10 clumps in the germplasm station. Data on measurable morphological characters were recorded in 2017/2018 planting season. Morphological traits were measured on ten random stems for each accession. A total of 37 morphological characters covering three main organs, namely leaves, stems, and buds, were evaluated according to Sastrowijono (1996). Most of these attributes were not subjected to selection in breeding programs; 7 of them related to leaves, 9 of them related to stem traits, while the other seven were subsidiary traits related to buds ( Table  2). All measures and observations were carried out in the ISRI's germplasm station and laboratory by means of metric rule and color chart, by the same operators for each attribute, considered stable enough observations for the different genotypes.

Molecular characterization
Deoxy Nucleic Acid (DNA) extraction was carried out using the CIMMYT Laboratory method. This method followed Saghai-Maroof et al. (1984) with minor modification by reducing one cycle for precipitation. DNA was precipitated without overnight incubation at room temperature but incubated for 60 minutes at -20ºC. This method had been used in the molecular research of sugarcane by Alix et al. (1998). The molecular characterization in this study used three types of microsatellite primers, as shown in Table 3.
The PCR results were visualized by electrophoresis on 1.5% agarose gel with 100 bp DNA ladder (Fermentas) and 1 kb DNA ladder (Promega) as DNA markers. The gel was visualized using Ethidium Bromide by incubated for 45 minutes in a dark room and observed with a UV transilluminator and documented with Doc XR Gel (Biorad).

Data analysis and interpretation
Morphological character data were transferred to the binary data format with number 1 for the character that appeared (present) and number 0 for the character that did not appear (absent). Likewise, molecular data in the form of amplified DNA bands were interpreted as qualitative data. Then the data were presented in the form of binary data as seen from the presence and absence of DNA bands. The presence of DNA bands was marked with the number 1, while the absence of DNA bands as indicated by the number 0 in the same array. The interpretation of this data referred to the research of Pan et al. (2003).The DNA bands generated through microsatellite markers were marked with numbers 1 and 0 and presented in the form of binary data.     Furthermore, binary data were also used to determine the value of genetic similarity and arrange genetic relationship trees. The genetic similarity coefficient used for morphological markers was the Simple Matching similarity coefficient (Sneath & Sokal, 1973). For molecular data, relationships between pairwise accessions were estimated using the Jaccard Coefficient Pan et al., 2003;Pocovi et al., 2020).
In this study, the accessions were then clustered by the Unweighted Pair-Group Method with Arithmetic Averages (UPGMA) cluster method with the Numerical Taxonomy and Multivariate Analysis System (NTSYSPC-2.02i) program (Schenk et al., 2004;Alwala et al., 2006;Parida et al., 2008;Pocovi et al., 2020). Cophenetics value matrices of the UPGMA clustering were used to test the goodness-of-fit of the clustering to the similarity matrix on which it was based by means of computing the product-moment correlation (r) with 1000 permutations (Mantel, 1967).

Morphological characterization
The 24 accessions of the Saccharum complex analyzed, based on 37 morphological characters, showed similarities and differences in morphological descriptors. The most similar morphological characters in the 24 accessions were on the leaves, internodes, and buds. On the leaves, similar traits are curved leaf blade, leaf color, auricle position, auricle height, dewlap color, the length of the hairy ligule, the width of the hairy ligule, hairy ligule position, midrib wax, removing leaf midribs, and leaf midrib color. On the internode, similar characters are internode color, internode shape, corky path, corky cracks, growth cracks, and the cross-section of the internode. In the bud, similar characters are bud furrow, bud furrow width, bud furrow length, growth ring position, growth ring shape, bud position, bud shape, the widest part of the bud, growth center, the hair on the edge of the bud basal, hair on top of bud, bud size and the edge of the bud wing.
The morphological characters distinguished between the accessions analyzed from the genus Erianthus and Saccharum in this study were located in the internode and bud. The distinctive character included the arrangement of internode (Erianthus: straight, Saccharum: zigzag); cross-section of internode (Erianthus: flat, Saccharum: round); the number of root buds (Erianthus: 1 row, Saccharum: 2-3 rows); and the widest part of the bud (Erianthus: under the middle of the bud, Saccharum in the middle of the bud). This resembles the observations reported by Amalraj and Balasundaram (2006), concerning the difference between Saccharum and Erianthus, among others, on the number of root buds. Erianthus had only 1 row of root bud number, whereas in Saccharum the number of root buds was two rows or more.
Qualitative data of morphological characterization were transformed into binary data to make a relationship dendrogram. Figure 1 shows that the phylogenetic relationship among 24 accessions of the Saccharum complex is divided into two large clusters at the similarity value of 0.66 (66%). The first cluster consists of 22 accessions and the second cluster consists of 2 accessions, namely PS 862 (Saccharum hybrid) and XOC55 (Miscanthus).
The greatest value of genetic similarity based on morphological characters was between IK 76-375 and IK 76-205 of 0.9683 (97%). The genetic distance between the two accessions was 0.0317, which means that the phylogenetic relationship between the two accessions is close. This was because IK 76-375 and IK 76-205 are both of the Erianthus genus originating from Kalimantan. In contrast, the lowest genetic similarity values were PS 862 and IJ 76-396 at 0.5714 (57%), and the genetic distance was 0.4286. PS 862 was a commercial sugarcane variety from ISRI's breeding program (Saccharum hybrid), while IJ 76-396 was Erianthus from Papua (Indonesia) so that the genetic relationship between the two accessions was farther than the other accessions. The clustering based on morphological markers (Figure 1) mainly corresponded to the classification at the conservation time. However, there were several accessions of Saccharum barberi, namely XOC 56 and XOC 55, which were distinct from one another, and accessions of Saccharum robustum, namely IJ76-412 and IJ76-415, which were not in the same group. Likewise, M 442-51 and PS 862 were not included in the Saccharum hybrids cluster. Meanwhile, the accessions of Erianthus, S. spontaneum, S. officinarum, and S. sinensis were clustered according to the classification at the time of conservation.

Molecular characterization
The results of DNA amplification of the three microsatellite primers are shown in Figure 2-4. The three primers produced 90 DNA bands with varying sizes of base pairs. In this study, the DNA bands analyzed were DNA bands that could be seen clearly, regardless of the difference in intensity (thickness or thinness of the DNA bands). Figure 2 shows that the microsatellite amplification of 24 accessions using mSSCIR5 primer resulted in 44 polymorphic DNA bands, with sizes ranging from 83 bp to 1600 bp. Meanwhile, Figure 3 shows that the SMC1047HA primer amplified the DNA of 24 Saccharum complex accessions to produce 22 bands. The sizes of the DNA bands ranged from 76 bp to 1642 bp. In Figure 4, it can be seen that the MSCA053C10 primer produced 24 DNA bands with sizes ranging from 98 bp to 950 bp.
Three pairs of microsatellite primers produced 90 DNA bands in the range of 76 bp -1642 bp. The data generated from the three microsatellite primers were transformed into binary data to form a dendrogram for genetic relationships using the NTSYSpc-2.02i program. Jaccard's genetic similarity coefficient used a scale of 0.00 -1.00. The value of 1.00 indicated 100% similarity, which meant that the value of genetic distance or dissimilarity was 0.00 (Indah et al., 2008;Pocovi et al., 2020).

. Hubungan kekerabatan 24 aksesi Saccharum complex berdasarkan penanda molekuler menggunakan koefisien Jaccard
The results of this study similar to the results of research conducted by Olieviera et al. (2017) to assess the genetic diversity of 26 accessions of sugarcane from the Active Germplasm Bank of the Embrapa Coastal Tablelands, using inter-simple sequence repeat (ISSR) molecular markers. Sixteen primers were used, resulting in 87 fragments with 91.13% of polymorphism. The similarity of the individuals ranged between 0.22 and 0.87. Four distinct clusters were formed using UPGMA. Despite the large distribution of accessions in most groups, indicating great genetic diversity, these were not separated by a group of the same species in each subcluster, which common alleles could explain among different species, since they are of the same genus. Similar results were found by Raj et al. (2016) that there was a close relationship among the different species that constitute the Saccharum complex; however S. spontaneum, contributing to greater diversity.
Subcluster I.2 consisted of the S. robustum accessions. Large cluster II at a similarity value of 0.04 was divided into two subclusters, namely, subgroup II.1 consisting of accessions of the genus Erianthus originating from Papua and subcluster II.2 consisting of (A) accessions from S. barberi; (B) accession of the genus Miscanthus; and (C) accession of the Saccharum hybrid. Figure 5 shows that the grouping based on the molecular markers of the 24 accessions analyzed were grouped according to species, this is the same as the grouping at the time of conservation. The genetic similarity value for the 24 accessions was very low at around 0.20.
As reported by Aitken et al. (2018), who analyzed the diversity of several accessions to S. spontaneum species from 21 different countries using molecular markers, the results of their analysis revealed that the majority of the accessions were clustered according to the country of origin. Genetic similarity ranged from 0.25 to 0.54, with the highest diversity in accessions collected in Indonesia, followed by China, India, and Thailand, and the lowest in the Philippines. Likewise, the results of the analysis conducted by Manish et al. (2014) on 40 genotypes of Indian sugarcane using microsatellites. The molecular results obtained in the study of Manish et al. (2014) are represented and agree with the possible evolutionary course of sugarcane genotypes. The UPGMA cluster analysis of 40 genotypes in that study produced meaningful grouping based on pedigree or geographical origin of the accessions. The grouping patterns of 40 genotypes are based on their genetic similarity pattern which shows that sugarcane genotypes from the same geographical regions tend to cluster together which may resulting from a shared evolutionary pathway.

Genetic relationship based on morphological and molecular markers
Information about the value of genetic similarity can be used as a reference to determine the value of genetic distance, phylogenetic relationships, and the level of genetic diversity. The value of genetic similarity is directly proportional to the genetic relationship but inversely proportional to the value of genetic distance or dissimilarity. The greater the value of similarity between two accessions, the greater the genetic relationship (closer), but conversely, the value of the genetic distance is small.
The markers used in this study were morphological and microsatellite markers, each of which had advantages and disadvantages. The use of these two markers in this study was to support and complement the analysis of the grouping and phylogenetic relationship of 24 accessions of the Saccharum complex. The data derived from morphological and molecular character were qualitative, so they were transformed into binary data before making the dendrogram of genetic relationship.
The dendrogram of the genetic relationship of the morphological characters was constructed based on the Simple Matching (SM) coefficient similarity (Figure 1). The dendrogram shows that the 24 accessions of the Saccharum complex including the genus Erianthus and Saccharum have a close relationship of 0.5714-0.9683 with a genetic distance of 0.0317-0.4286. This low diversity was due to the 24 accessions analyzed with similar morphological features from the 37 characters observed. Therefore, grouping based on morphological markers was considered insufficient to verify the classification of germplasm collections managed by ISRI, so that molecular markers were still needed to support the validity of existing groupings.
On the other hand, the value of genetic similarity based on molecular markers between 24 accessions ranged from 0.000 to 0.889 ( Figure  5). Duplication (similarity value = 1.00) was not found among the 24 clones tested. A high genetic similarity was observed in some accessions because the accessions were included in groups of the same species.
Qualitative data on the characterization of molecular markers were used to construct a dendrogram of genetic relationship with Jaccard's similarity coefficient. The Jaccard's coefficient is appropriate for molecular markers because Jaccard's coefficient only considers the characters or bands that appeared (present) as a contributing factor to individual similarity and ignores zero (0) or absence. As in the research of Cordeiro et al. (2003), Widyasari et al. (2008) and Pocovi et al. (2020) used Jaccard's coefficient to analyze the genetic diversity of sugarcane which based on molecular markers.
The grouping based on the molecular data shows that of the 24 accessions evaluated, each accession was grouped according to its species. This means that based on molecular markers, the grouping of ISRI's sugarcane germplasm collections was still in accordance with the grouping at the conservation time.
The results of this study indicate that there are differences in the value of genetic similarity in the 24 accessions analyzed based on the markers used. According to morphological markers, the value of genetic similarity was higher than the value of genetic similarity based on molecular markers. The morphological markers caused the difference because the observed morphological characters were 37 characters with 126 descriptions, while the molecular markers, using three microsatellite primers, could produce 90 DNA bands. In addition, descriptions of morphological characters that were equally absent were considered the same variable, even though their morphological characters were different. It was a cause of increased genetic similarity value when the genetic relationship was analyzed using morphological markers.
Another factor influencing these differences is the difference in the similarity coefficient used on morphological and molecular markers in evaluating the genetic relationship of those 24 accessions. Morphological markers used the Simple Matching similarity coefficient, which involved the absent value (d, 0-0). In contrast, the molecular markers used the Jaccard's similarity coefficient, which only involved the present values (a, b, c).
However, the two markers gave the same grouping results. This can be seen in the accessions of S. officinarum, TD50 with TD53 into one group; S. spontaneum, IK76-36 and IK76-41 in one group; S. sinense, 15OC18 and 15OC19 in one group; Erianthus clones originating from Irian Jaya (IJ) also form a separate subgroup of Erianthus originating from Kalimantan and Sulawesi. Meanwhile, the classification of several accessions was different based on morphological and molecular markers. It was shown in the accessions of S. barberi, S. robustum, and Saccharum hybrid. The fact that grouping based on molecular markers is not exactly the same as a grouping based on morphological markers is also found in Pocovi et al. (2020).
Regarding the accuracy of genetic relationships, some researchers argue that molecular markers are considered to be more accurate in phylogenetic analysis than morphological markers. This is because molecular markers analyze individuals at the DNA level, whereas morphological markers had several drawbacks that could affect the accuracy of the data. Bagali et al. (2010) stated that morphological markers are markers that recognize a character according to their natural appearance based on the observer's view (subjectivity). In addition, Martinez et al. (2003) reported that morphological markers were relatively insignificant and inefficient for accuracy in distinguishing closely related genotypes and analyzing their genetic similarities when compared to DNA fingerprinting techniques. However, morphological markers are helpful at the beginning or early stage, fast, simple, and inexpensive to identify varieties, and could be used as a general approach to measuring genetic diversity between cultivars of different phenotypes. Karuri et al. (2010) stated that it was necessary to combine them with genetic-based markers due to limitations in morphological markers.

Conclusion
The grouping of the 24 accessions of the ISRI's Saccharum complex germplasm collection was based on molecular markers according to the grouping at the t conservation time. Meanwhile, the grouping based on morphological markers in several accessions did not match the classification at the conservation time. This occurs in accessions of the Saccharum robustum, Saccharum barberi, Saccharum hybrid, and the genus Miscanthus. Based on molecular markers, the 24 accessions of the Saccharum complex had a high degree of diversity. These clones can be used as a combination of crossing parents to produce new superior sugarcane varieties.