Lablab is a very familiar genera of the family Leguminosae, cultivated and consumed primarily for the seeds and pods. The genera is monotypic containing a single species Lablab purpureus (L.) Sweet which further has two subspecies viz. Lablab purpureus bengalensis (Jacq.) Verdc. and Lablab purpureus pupureus (L.) Verdc. Lablab bean have immense potential due to their inherent ability of withstanding drought, soil acidity and tolerance against mineral toxicity such as Aluminium toxicity (Ansari et al., 2019), besides they are also important as catch crops and herbage (Afsan and Roy, 2019). The genus is extremely diverse and about 3000 accessions of germplasm have been reported from the different regions of the world (Maass, 2010). The Indian sub-continent boasts of a large number of indigenous collections across the country (ICAR- NBPGR National Gene Bank). The north eastern region of India, Assam in particular is a remarkable natural gene repository with several indigenous germplasm traversing generations and their farmlands. In a study by Sarma et al., (2010) nine (9) landraces of Lablab has been reported from the different areas of Assam and Northeast.   
       
Pods of Lablab beans are mostly consumed as vegetable (Dwivedi et al., 2023), besides dried seeds are used as an alternative to common pulses and it has a great potential in overcoming problems related to malnutrition as it is nutritionally very rich (Deka and Sarkar, 1990). The improvement of this underutilised crop is important; however, the self-pollinating nature of the crop often poses as a hindrance in generating variability and selection is one of the ways to create variability. Hence it is imperative to understand the important traits for the improvement of the crop. Lablab has been an integral crop in the homesteads and farmlands in Assam, but these genotypes have not made itself to the core collection and are less utilised in crop improvement programmes. In this regard Principal component analysis (PCA) or canonical root analysis serves as an important analytical tool which can help us to mark out certain traits that can be used for plant selection. Natural variations, their exploration and the extent of variation could be useful for genetic improvement of germplasm (Ullah et al., 2022). Also, the selection of characters will be easier and more precise by understanding the correlation between the traits. The present study will be helpful in understanding the relationship among the traits and also in finding out the suitable genitors for further breeding programme.

Collection and planting of plant material
 
Germplasm of 16 (sixteen) different genotypes were collected from the various locations of the six agro-climatic zones of Assam (Fig 1) and maintained in the departmental garden, Department of Botany, Cotton University, Assam from 2020 to 2023. The average weather parametres of the study site has been represented in the table (Table 1) (https://mausam.imd.gov.in). The experiments were laid in a randomized block design (RBD) with three (3) replications and 1 x 1 meter spacing. Quantitative characters as well as the stages of scoring are indicated in table (Table 2). Measurement of length and width was carried out using a calibrated scale and vernier caliper. The individual traits were recorded in 3 (three) replicates and the parameters were observed in 4 (four) consecutive growing seasons.






 
Statistical analysis
 
The data were analysed using IBM SPSS Statistics v.28. Genotypic variance (σ²_g  ), Phenotypic variance (σ²_p), Environmental variance (σ²_e), Genotypic coefficient of variation (GCV), Phenotypic coefficient of variation (PCV), were calculated based on method by Burton and Devane (1953). Heritability in broad sense (h²_b), Genetic Advance (GA) and Genetic Advance Percent of Mean (GAPM) were estimated using the formula suggested by Lush (1949) and Johnson et al. (1955). The principal component and Eigen values for the individual data and the component scores for the various data were determined (Pearson, 1901; Hotelling, 1933). The linear correlation between the variables were determined at 5% level of significance (Pearson, 1896). The genetic divergence was measured using Mahalanobis D² statistics (Mahalanobis, 1936) and cluster analysis was carried out by Tocher’s Method (Rao, 1952).

Summary statistics for variables
 
Descriptive statistics were measured for sixteen (16) characters (Table 3) and the characters considered are shown in the Table 2. The highest variation was found for seed/plant with a CV of 55.45% and lowest was 5.69% for dal recovery.


 
Assessment of variability
 
The estimation of genotypic coefficient of variation (GCV), phenotypic coefficient of variation (PCV), heritability, genetic advance are shown in Table 4. The mean sum of square (MS) is significant for most of the characters which revealed significant variations among the genotypes, except for the traits 100 seed (dry) weight and dal recovery. Lowest GCV and PCV was found for seed yield /plant and was recorded as 5.67 and 5.74 respectively. While, highest GCV and PCV was recorded as 52.97 and 56.14 for average pod (fresh) weight. The results indicate the influence of environment on the traits. However, less difference between GCV and PCV indicates more influence of genetic component in phenotypic expression of the traits rather than environment.  High GCV and PCV value for the traits except for leaf length (17.00 and 17.28), leaflet length (17.59 and 18.95), raceme/ plant (17.45 and 17.49), 100 dry seed weight (10.16gm and 18.30 gm) and dal recovery (10.16 and 18.30) indicate high variability among the genotypes for the traits that provides scope for selection. Seed yield/plant showed low but significant GCV and PCV (5.67 and 6.74). Heritability insights the genetic basis of the traits of a population, thus provides scope for efficient selection of traits from diverse genotypes available. High heritability (>60) was recorded for most of the traits except dal recovery and pod/raceme (30.8%). Highest heritability was recorded for pods per plant (99.9%) followed closely by traits like seed yield/plant (99.6%) and days to first flowering (99.6%). Genetic advance (GA) and Genetic advance percentage of mean (GAPM) are related to prediction of the expected results of a selection or the advancement which a particular trait is likely to achieve. However, GA represents the absolute value and GAPM represents the relative value based on the population mean. High genetic advance (GA) for the traits under investigation was recorded for raceme/plant (27.32%) pod width (80.64%), average pod (Fresh) weight (244.43%), pod/raceme (53.13%) and pod/plant (21.87%) and that indicates presence of wide of genetic alleles which primarily control the traits with additive effect, thus provide advantage for selection.  All other traits which exhibited low genetic advance may be due to the control of both additive and non-additive genes in combination over the traits. The values of GAPM were found to be high for 13 (thirteen) traits and only 3 (three) traits showed moderate GAPM, as proposed by Johnson et al., (1955). Highest GAPM was recorded for fresh average pod weight (102.94%) and the lowest was for the trait seed yield/plant (11.54%). None of the traits falls under the low GAPM category. Abundant diversity with high heritability and genetic advance in target germplasm is indispensable for breeding programme (Rasheed et al., 2023). Influence of genetic factor on the traits, high heritability and genetic advance provide the scope for divergent parental line selection and heterozygosity advantage for the yield attributing traits in groundnut (Shekhawat et al., 2023).

​
 
Pearson’s correlation-coefficient analysis
 
Pearson’s coefficient efficiently measures the strength and linear correlation relationship concerning two variables. In the present study (Table 5) strong positive significant correlation was seen for leaf width with leaflet length (0.898), pod length with fresh pod weight (0.709), fresh seed weight with dry seed weight (0.863), 100 fresh seed weight with seed yield/plant (0.807), dry seed weight with seed yield (0.806), pod/raceme with pod/plant (0.730). Similarly positive significant correlation was also seen for leaf length with leaf width (0.623), leaflet length with leaf length (0.661) and raceme/plant (0.604), pod length with 100 dry seed weight (0.633), pod weight with 100 dry and fresh seed weight (0.691 and 0.694). The locule/pod has shown perfect positive significant correlation with seeds/pod (1.000). Moderate correlation was seen for the characters raceme/plant with pod/plant (0.501), leaf width with pod width (0.506) and100 seed weight with dal recovery (0.530). Pod/raceme, pod length, seed/pod, seed weight showed direct correlation on total yield of the plant. The study is in consent with the findings of Girgel et al. (2021) and Singh et al. (2018) who reported strong positive correlation among the yield attributing traits in Phaseolus vulgaris L. and Carica papaya L. respectively. Significant positive correlation among the traits indicates the scope for selection and to improve multiple traits simultaneously.

​
 
Principal component analysis (PCA)
 
PCA is an important approach to find out the total variation in a population and interrelationship among the variables, thus plays an important role for selection of traits or germplasm for genetic improvement. The PCA under the present investigation was considered for the variables with Eigen values more than 1(one) as per Kaiser Rule (1961). The Eigen value, variability % and cumulative % values are indicated in the table (Table 6). Highest variability was shown by PC1 (32.95%) followed by PC2, PC3, PC4 and PC5 which represented 20.33%, 15.60%, 8.63%, and 7.30% respectively (Table 6). The total variance of 5 (five) components was recorded as 84.81%. Screw plot of Eigen value based on 16 traits has been represented in Fig 2. In PC1 all the characters except raceme/plant contributed positive loading value. The Scree plot shows that from the 11^th PC (PC11) there was very little variance and the graph was more or less linear.




       
The positive and negative loadings of the various variable are represented in the Rotated Component Matrix (Table 7, Table 8) and PCA biplot (Fig 3). For traits located at narrow, wide and right angles the relationship was considered as positive, negative and no relationship respectively (Mohanlal et al., 2023). The PCA biplot of the present study shows positive relationship between the traits like 100 seeds wt. (Fresh) and seed yield/plant, dal recovery and fresh average pod weight, pod width and leaf length, leaf width and leaflet length, locule/pod and seeds/pod.






       
In PC1 most of the yield attributing traits like 100 dry seeds wt., 100 Fresh seeds wt., seed yield/plant, fresh average pod weight, dal recovery, pod length, leaflet length, pods per plant and leaf width exert positive loadings but the trait raceme/plant showed negative relationship (Table 7). In PC2 positive contributions were made by leaf length (0.865), raceme/plant (0.818), leaflet length (0.700), leaf width (0.673), dal recovery (0.403), pods/ plant (0.288), fresh average pod weight (0.233), pod width (0.220), locule/pod and seeds per pod (0.163), seed yield/plant (0.131). While negative contribution was seen for the traits like 100 seeds wt. (Fresh) (-0.024), pod length (-0.089), 100 seeds wt. (Dry) (-0.104), pods per raceme (-0.191), days to first flowering (-0.585).
       
Traits like pods/raceme (0.915), pods/plant (0.861), seed yield/plant (0.283), 100 seeds wt. (Fresh) (0.268), raceme/plant (0.132), leaf length (0.053) showed positive contribution in PC3. In case of PC4, strong positive effects were contributed by locule/pod and seeds/pod (0.942) followed by leaf width (0.299), pod length (0.284), leaflet length (0.278), pod width (0.198), leaf length (0.160), 100 seeds wt. (Fresh) [0.110], 100 seeds wt. (Dry) [0.098], seed yield/plant (0.074) and fresh average pod weight (0.039). Moreover strong negative contribution was found for some traits.  In the  PC5, traits like pod width (0.871), fresh average pod weight (0.450), pod length (0.263), leaf width (0.261), leaflet length (0.236), Pods per raceme (0.181), locule/pod and seeds/pod (0.117), leaf length (0.085), dal recovery (0.083), 100 seeds wt. (Fresh) [0.069], 100 Seeds wt. (Dry) [0.007] .The proportion of variance decreases from PC1 to PC5. The study showed high percentage of variability i.e. 84.80% by the traits under PC1 to PC5. PC1 constitutes majority of variability accounting to 32.95% of the total variability. The significance of the characters in principal components has been reported for physiological and biochemical parameters of rice cultivars by Chunthaburee et al. (2016) and the legume Vigna radiata (L.) Wilczek by John and Aravinth (2024). In the study most of the yield attributing traits are in PC1, showed high relationship among themselves (Fig 3) and may be considered for selection.
 
Genetic divergence and cluster analysis
 
The cluster analysis of genotypes revealed four (4) clusters. The first cluster is the largest comprising of 13 (Thirteen) genotypes viz. CUCYT22001, CUCYT22002, CU CYT2 2003, CUCYT22004, CUCYT22005, CUCYT22008, CUCYT2 2009, CUCYT22010, CUCYT22011, CUCYT22012, CUCY T22013, CUCYT22014, CUCYT22015 (Table 9). The other three clusters  include only one genotype each ie. CUCYT22006, CUCYT22007 and CUCYT22016 in the second, third and fourth cluster respectively (Table 9). The maximum inter-clusteral distance was observed between cluster 2 and cluster 4 with a distance of 11283.13 and minimum was between cluster 1 and cluster 4 with a distance of 3112.670 (Table 10 and 11; Fig 4). The occurrence of larger inter-cluster distance is indicative of the occurrence of larger genetic diversity. Intra-cluster distance, the maximum was recorded for cluster 1 (952.51) indicates the relatedness of traits. The occurrence of rich genetic diversity was highlighted by Prasanna et al. (2023) in Vigna mungo (L.) Hepper using distance analysis. Genetic diversity in a population is the prime requirement for the plant improvement program (Appalaswamy and Reddy, 2004) that can ensure food security.

The current scenario for attaining food security necessitates the utilisation of the existing genetic diversity. Under the present investigation, less difference between PCV and GCV indicates the influence of genetic factor on the phenotypic expression of the certain traits. Heritability determines the population structure and transmissibility of the traits. High heritability and high genetic advance for certain traits signifies the variability among the traits is due to genetic factor rather than the environment. High genetic advance for the traits favours for effective selection for substantial improvement of germplasm. Significant positive correlation among the traits under Pearson’s Correlation- coefficient analysis indicates the scope for selection and to improve multiple traits simultaneously. On the other hand, negative correlation for certain traits reflects adverse effect towards selection.
       
The understanding of the correlation among the traits can help in accentuating a particular trait and thereby help in the proper selection of the agronomically significant trait. Lablab has high potential as a futuristic crop though highly underrated; hence high correlation among yield and yield attributing variables observed in the study significantly indicates the advantage for targeted selection. From the study it can be concluded that traits like fresh average pod weight, 100 seeds weight (fresh), pods per raceme, seeds per pod, locule per pod, pod width and pod length may contribute in making Lablab adaptable to diverse environment, generating variability and subsequent genetic gain. The segregation of genotypes into different clusters indicates the diversity existing among the genotypes of Assam, which needs to be harnessed before it cease to exist. However, the differential expression of the traits in the different genotypes need to be ascertained by analysis of expression pattern of the genes. This can further be supplemented by the analysis of the differential expression of the traits as per the edaphic and climatic conditions to develop a climate smart crop.
 
Funding declaration
 
This research has not received any grant.
 
Disclaimers
 
The views and conclusions expressed in this article are solely those of the authors and do not represent the views of their affiliated institutions. The authors are responsible for the accuracy and completeness of the information provided, but are not liable for any direct or indirect losses resulting from the use of this content.
 
Informed consent
 
Not applicable.

I, on behalf of all the authors of the manuscript do hereby declare that all the authors have no conflict of interest to declare. All co-authors have seen and agreed with the contents of the manuscript and there is no financial interest to report or involvement in any organization or entity with any financial interest while carrying out the work.