Assessment of morphological characteristics among upland rice ( Oryza sativa and Oryza glaberrima ) germplasm

Rice is an important staple food crop that feeds over half of the global population and has become the cereal that provides a major source of calories for the urban and rural poor in Africa. This work aimed to evaluate the morphological of rice (Oryza sativa and Oryza glaberrima) germplasm. In the present study, 14 quantitative traits were used across 48 accessions or genotypes obtained from Central Agricultural Research Institute (CARI), Liberia and Plant Genetic Resources Research Institute (PGRRI), Ghana. In this research, Completely Randomized Design was conducted to study the genetic variability among the 48 genotypes or accessions obtained from CARI, Liberia and PGRRI, Ghana. Field data taken included 14 quantitative traits scored using the IRRI descriptor list. Analysis of variance revealed highly significant difference (P≤ 0.01) among the accessions for all quantitative traits studied. Four significant principal components analysis were identified and accounted for 55.3%. PC1 had Eigen-value of 18.5%, whereas PC2 accounted for Eigen-value of 14.5%. PC3 contributed 12.2% whereas PC4 had 10.1%. Correlation analysis indicated the length of ligule was highly significant and positive leaf width blade. Similar observation was made with grain length and length of ligule. Some accessions in the biplot showed longer vector distances, while shorter vector distances were observed referencing PC1; Gh1578 recorded the longest vector distances while Gh1526 and LAC 23-1, recorded the moderate distances from the vector origin at the similarity coefficients at 90%. The highly distant genetic diversity was found between ACSS37 and ACSS1 from Ghana and Liberia. Cluster X was the largest of all the clusters while Clusters VII and VIII were the second largest clusters with seven accessions each. The outcome of this study should be useful for the management of the germplasm conservation and future rice genetic improvement. However, all the accessions may be cultivated over time at different locations on the field to ascertain their stability and purity.


INTRODUCTION
Rice (Oryza sativa L., 2n = 24), a member of Poaceae (Gramineae) is the world's most important staple food crop that feeds over half of the global population (Ishimaru et al., 2017).It is cultivated in tropical and subtropical regions.Rice is grown in more than 114 countries, over an area of 161.4 m ha in a wide range of ecosystems under varying temperature and water regimes with the production of 466.7 mt (on milled basis) (FAO, 2011).According to Ansah et al. (2017), approximately 20 million farmers are engaged in rice production in sub-Sahara Africa (SSA) and about 100 million people depend on it directly for their livelihoods on the continent.
Rice is rapidly becoming a staple food in the African diet; and its production in SSA continues to be outpaced by consumption as a result of low and stagnated production.Imported rice accounts for 50% of sub-Saharan Africa's rice requirement (Notarnicola et al., 2017).Rice is no longer a luxury food but has become the cereal that constitutes a major source of calories for the urban and rural poor.Rice production in SSA has been bedeviled with conditions such as environmental degradation due to pesticide usage, excessive water usage, and nutrient contamination, methane emission and ammonia volatilization and these conditions require urgent attention (Luther et al., 2017).
A wide range of technologies are available and can be used as tools for reducing these adverse consequences of rice production; but they are, however, not extended to majority of rice growers or farmers (Notarnicola et al., 2017).Self-sufficiency in rice production is, however, declining as demand increases.Little attention has been paid to the improvement of Liberian and Ghanaian rice germplasm evaluation and the genetics of some quality traits.Thus, there is very little information available on the genetic diversity of Liberian and Ghanaian rice germplasm for crop improvement and conservation purposes.There is an urgent need to increase and improve the production of rice in Africa in order to meet up with the high demand.The need for increasing rice cultivation depends not only on cultural/traditional practices, but also, on their inbuilt genetic potential to withstand stresses.A successful breeding programme will depend on the genetic variability of a crop for achieving the goals of improving the crop and producing high yielding varieties (Onyia et al., 2017).The first step in achieving this is to evaluate and characterize available rice germplasm or genotypes at the morphological stage.

METHODOLOGY Planting materials and experimental design
Forty-eight (48) upland rice accessions were evaluated from Ghana and Liberia.Thirty-five Ghanaian genotypes were obtained from Plant Genetic Resources Research Institute (PGRRI) at Bunso Ghana, while 13 Liberian accessions were obtained from the Central Agricultural Research Institute (CARI), Suakoko in Bong County, Liberia.Experiment was carried out at the Insectary Laboratory, Faculty of Agriculture, Kwame Nkrumah University of Science and Technology (KNUST) and pots were arranged in Luther et al. 907 complete randomized design with three replicates.The compound fertilizer of NPK (15-15-15) was applied by ring method seven days after planting at the rate of 1.5 g in each pot as a first dose and second dose of NPK was applied one month after the first at the same rate per pot.The third dose was the urea fertilizer.The urea was the third fertilizer applied in the maximum tillering stage one month after the second dose of NPK.Pesticide Lambda Master 2.5 EC (25 g lambda-cyhalothrin/liter) at a dosage of 100 mL/15 L of water (600 mls/ha) were used for pest outbreak.Manual weeding and birds watching were done at tillering and reproductive stages using bird net till harvesting time.

Data collection
Evaluation of the rice accessions was carried out for different morphological parameters representing the vegetative growth stage of rice.Trait selection and measurement techniques were based on IRRI standard evaluation system of rice (Hien et al., 2007).Culm diameter at basal internode (Culm diameter (mm) was measured using vernier calipers and the mean was computed), leaf length of blade (length of the topmost leaf blade below the flag leaf on the main culm in centimeters), leaf width (width of the widest portion of the blade on the leaf below the flag leaf in centimeters), flag leaf length width (leaf length was measured from the base to the tip of the flag leaf, rounded off to the nearest millimetre, while the width was measured at the widest part of the flag leaf and recorded to the nearest mm), and panicle number per plant (total number of tillers) were counted and recorded at the maturity stage before harvest), ligule length (length from the base of the collar to the tip of the ligule in millimeters), plant height (plant height (cm) was measured from soil surface to tip of the plant at reproductive stage using the measuring tape), productive tillers per plant (productive tillers/plant was obtained by counting the number of tillers per plant and averaged across replications for each accessions during the maturity stage).awn length (mm) (the awn is a long slender extension of the lemma in rice) was also measured.It was measured from the tip of the spike to the tip of the longest awn), panicle length of main axis (the panicle length of main axis was measured from the base of the panicle to the tip of the lemma or palea using the measuring tape), one hundred grain weight (one hundred well developed seeds were randomly selected per replication for each accession.The seeds were obtained from the harvested samples of accessions after harvest; dried to 13% moisture content and weighed on a balanced precision scale (METTER PM 400) to determine the 100 grain weight, grain length (the grain length was measured as the distance from the base of the lowermost glume to the tip (apiculus) of the fertile lemma or palea), and grain width (to obtain the grain width, it was measured as the distance across the fertile lemma and palea at the widest point using the callipers at post-harvest stage).

Data analysis
All recorded agro-morphological traits were analyzed using two complementary procedures; Microsoft Excel was used to record and organize the data.The quantitative data were subjected to Analysis of Variance (ANOVA) using the GenStat Statistical package version (12 th edition, VSN international, Hemel Hempstead) to calculate the relationship between the traits of the genotypes.

RESULTS
Fourteen quantitative traits were evaluated against 48 genotypes of rice received from Liberia and Ghana (Table 1).The evaluation of morphological traits usually reveals important traits which are essential to characterize the genetic resources.The mean, standard error, range, coefficient of variation, standard error of deviation, F probability of the least significant difference at 5% were computerized for 14 quantitative characters are shown in Table 2.The 14 quantitative traits were performed to determine the relative contribution on different traits to the total variation in rice.Among the 14 quantitative traits studied, productive tillers, grain width, sterile lemma length and grain length had variations of 42.0, 41.6, 28.1 and 24.2 (Table 2).On the other hand, leaf blade length, and flag leaf obtained the lowest variation among the traits (Table 2).
The principal components analysis based on the 14 quantitative traits was performed individually to determine the relative contribution of the different traits to the total variation in rice.The PCA is an ordination of multivariate technique that allows the use of biplots to visualize the relationship between the accessions and measured traits.
Pearson correlation among the 14 quantitative traits were highly significant as positive correlation was observed between length of ligule and leaf width of blade (r =0.93).Relationship between sterile lemma length and length of ligule was highly positively correlated (r = 0.92) (Table 4).Plant height was significantly positively correlated (r = 0.69).Grain length and length of ligule is highly significantly positively correlated (r =0.99) (Table 4).However, correlation between culm diameter at basal internode and length of ligule, sterile lemma length and culm diameter at basal internode, 100 grain weight of fully developed grain and culm diameter at basal internode, plant height and grain length were positively correlated (Table 4).Correlation between awn length and culm diameter at basal internode, panicle length of the main axis and length of ligule, plant height and 100 grain weight of fully developed grain were negatively correlated (Table 4).
The principal components analysis based on PCA biplot of the rice accessions revealed diverse grouping pattern among the 14 quantitative traits (Figure 1).While some accessions showed longer vector distances, shorter vector distances were observed for others.For PC1, Gh1578 recorded the longest vector distances.Gh1526 and LAC 23-1, recorded the moderate distances from the vector origin while GH3623, GH1588 and LAC23-2-3 recorded the shortest.GH1570 and GH1550 had the longest distances for PC2; whereas GH5173 and LAC 23-12 had similar vector in PC2 and were separated from the rest of the accessions (Figure 1).
A dendrogram was constructed for the 48 rice genotypes based on their morphological characteristics.Figure 2 shows that at a similarity index of 68%, the accessions clustered into 7 main clusters.The most distant genotype were Gh1540 and LAC23-1, which were found in the first and last of position of the dendrogram.Cluster I was the largest of all the clusters and contained 20 accessions with five sub-clusters.Cluster III was the second largest cluster with 16 accessions including two sub-clusters and Cluster IV had 8 accessions with two sub-clusters respectively.Cluster VI had three accessions with two sub-clusters and next was Cluster II which had two accessions only.Clusters V and VII had one accession each, which were LAC23-9 and GH1540.Cluster VI had three accessions and comprised two subclusters.Accessions in the same cluster have the same morphological characteristics and sub-clusters indicate that the accessions have some distinct traits from other members of the clusters.The 48 rice accessions (from Liberia and Ghana) showed no distinctive morphological characteristics based on geographical origin, as the analysis showed no group of accessions from either of the geographical locations divergently clustered.

DISCUSSION
In the present study, a set of rice genotypes from Liberia and Ghana were subjected to diversity analysis based on variation in morpho-phenological traits.To meet the future rice demand in Liberia and Ghana with the increasing population, one option is to increase the productivity per unit area of the land, thus the identification of more yield related agro-morphological characters is very much important.Plant height in rice is   et al. ( 2017).Tiller is one of the main attributing plant traits as indicated by Abbasi et al. (2015).Based on Table 2 statistical data analyzed, there was high significant difference of P< 0.011.The coefficient of variation and standard deviation recorded next 42.0%and 3.53 respectively.The accessions had a great variability with a high range (3-29) for number of productive tillers (Table 2).Better tillering capacity is a desirable feature to upgrade the yield potential of upland varieties.Ray et al. (2016) generally indicated that when rainfall is plentiful and the soil has good water-retention capacity, the high-tillering and short-statured varieties definitely respond better to nitrogen and yield higher than do the taller types.The accessions that produced more productive tillers will contribute to increased yield in a breeding program and could be selected as base genotypes for further improvement.In breeding applications, according to Chen et al. (2014), grain size is usually evaluated by the grain weight, which is positively correlated with several characters including grain length, grain width and grain thickness.It is a major determinant of grain weight; one of the three components (number of panicles per plant, number of grain per panicle and grain weight of grain yield).The grain length ranged from 0.96 to 2.08 cm and width 1.16 to 3.10 cm (Table 2), thus inferring rice accessions were largely long-grain.Although the preference for rice grain characteristics varies with consumer groups, long and slender grains are generally preferred and are good valuable attributes that could be exploited to improve the grain characteristics of local rice accessions (Cuevas et al., 2016).Similar variability were reported by Javed et al. (2015) who studied 12 genotypes of coarse rice to check their yield performance in Kallar tract and reported highly significant variation for different traits.This variation in the grain yield might be due to the environment and genetic constitution of accessions (Xie et al., 2015) or the correlation of grain yield per plant with various yield contributing characteristics such as; fertility of soil, flag leaf area, number of grains per panicle and grain weight which showed positive correlations.Similarly, Jones et al. (2015) reported positive correlation among number of panicles per plant, panicle length, number of grains per panicle, 100-grain weight and grain yield per plant-type.The grain shape character also showed the highest variation in studies conducted in Pakistan by Siddiqui et al. (2007).Core collection is important in germplasm characterization.Accessions selected for this study were 48 in total from Liberia and Ghana.Among the accessions studied, 18 out of the 48 accessions were distant from the rest, and were selected to constitute a core collection for further improvement.The concept of a core collection was introduced by Krueger et al. (2015) with the intent of using the core collection to minimize the cost of germplasm conservation while ensuring maximum genetic diversity.
According to Wijayawardhana et al. (2015), cluster analysis has the singular efficacy and ability to identify crop accessions with highest level of similarity.The dendrogram obtained from the present study also proved the above statement in terms of similarity existing among accession further.Baloch et al. (2016) also proved that agro-morphological traits can be used effectively to characterize the rice cultivars.Wangpan et al. (2018) have reported a similar variability of rice varieties.Results of the present study have shown similarities to the findings of Wangpan et al. (2018) in terms of dendrogram analysis, clustering groups and PCs analysis with some exceptions.These exceptions were identified in terms of variations in cluster formation and grouping behaviors of their tested rice varieties.Further, these exceptions are possible due to the variations in external conditions such as soil types, soil fertility levels (Murikov et al., 2017) and soil moisture regimes (Okii et al., 2014) associated with two cropping systems.Furthermore, the genetic make-up of seed, environment and field management practices has been reported to influence the morphology of a crop (Xiong et al., 2018).Therefore, the identification of agromorphological characters such as number of tillers, plant height, grain length, grain width and flag leaf that are important to change the rice crop architecture which have greater implication in this attempt.Hence, the data from the current study with other agro-morphological data may be widely applicable in future rice crop improvement programs.

Conclusions
Results of the present experiment have clearly indicated the importance of agro-morphological traits to identify naturally existing distinguishable clusters.With the use of this information, plant breeders can effectively select morphologically more distinct individuals for their breeding programs.Morphological traits are used as a preliminary evaluation tool due to their easiness and can be employed as a common approach for assessing genetic variability among phenotypically distinguishable rice accessions.This study highly focused on the reproductive characters of the rice plant irrespective of the post-harvest characters.Most of the previous studies have reported the morphological variability in favor of both reproductive and post-harvest characters.Hence, the present study interprets a considerable amount of morphological diversity along with the reproductive traits of the rice plant.Therefore, rice reproductive traits can be used effectively in order to capture a considerable morphological variability associated with rice germplasm.

Figure 2 .
Figure 2. Dendrogram generated based on the 14 traits of the forty-eight (48) rice genotypes using UPGMA.
Least Significant Difference (LSD) at 5% was used to separate the treatments means.MINITAB ® 17 Statistical Software was used to perform the principal component analysis (PCA).
Tel: +231+886-462-538.Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License

Table 1 .
Accessions of rice's, their sources and collection countries.

Table 2 .
Summary statistics of 14 quantitative traits measured on 48 rice accessions from Liberia and Ghana.

Table 3 .
Principal components analysis based on the 14 quantitative traits.
, 2018).Tall plant type is very typical of landrace genotypes which exceed in their capacity to support panicle growth by large stem reserve mobilization.Ali et  al. (2008)observed relatively greater range in plant height than the other characters.The smallest plant height was recorded for accession GH1571 cm and accession GH1550 cm recorded the highest value.This was true because the semi dwarf plant type was extensively utilized in the rice (O.sativa) cultivars throughout the world.However, depending on the part of the world with improvement in farmers' lives, there is a growing desire to combine desirable characteristics of tall varieties' with yielding ability and a new type of architecture: intermediate plant height as stated by Zafar

Table 4 .
Pearson correlation coefficients among the quantitative traits studied.WOB=Leaf width of blade; LOL=Leaf length of ligule; FLL= Flag leaf length; FLW=Flag leaf width; DABI=Culm diameter at basal internode; PAN=panicle number per plant; SLL=Sterile lemma length; AL= Awn length; LOMA= Panicle length of the main axis; WOFGD=100 grain weight of fully developed grain; GL= Grain length; GW=grain weight; PH= plant height.