Unraveling genetic divergence using principal component analysis for yield and fibre quality traits in cotton genotypes

Introduction

Cotton (Gossypium spp.) is a cornerstone of the global textile industry and a critical commercial crop for India. As one of the oldest cultivated plants, it's primarily grown for its fibre and oil in tropical and subtropical regions. India plays a significant role in global cotton production, contributing approximately 20.67% (ICAR-AICRP (Cotton)-Annual Report 2024-25) [5] . It is crucial for increasing the country's economy and is popularly known as "white gold" (Komala et al. 2018) [7] . Belonging to the genus Gossypium within the Malvaceae family, cotton exhibits extensive genetic diversity. This genus comprises 50 known species: four cultivated, 44 wild diploids, and two wild tetraploids (Percival and Kohel, 1990) [11] . Globally, four main cotton species are cultivated, including two diploids (Gossypium herbaceum and G. arboreum) and two New World tetraploids (G. hirsutum and G. barbadense). However, the genetic base of cultivated cotton genotypes remains narrow (McCarty et al. 2004; Shakeel et al. 2015) [9,14] . To broaden this genetic base, especially through hybrid breeding programs, an understanding of genetic divergence among available germplasm is essential. Genetic variation among cotton genotypes with respect to morphological and fiber quality traits has been studied to facilitate their improvement (Ruilian et al. 2005) [8] . Researchers have employed principal component analysis (PCA) to assess similarity among genotypes based on these traits and to classify them (Sarwar et al. 2021) [13] . Principal Component Analysis (PCA) and cluster analysis are widely utilized to explore genetic similarities and differences across genotypes, grouping them into distinct clusters (Naik et al. 2016) [10] . PCA is particularly effective at partitioning total variation, making it highly suitable for identifying elite parental lines by highlighting the key traits that drive divergence. It also aids in estimating phenotypic variability and genetic diversity among existing genotypes supporting informed selection decisions.

By revealing which traits contribute most to variability, PCA provides essential insights for choosing genotypes with both agronomic and genetic significance (Isong et al. 2017) [6] . The main objective of the present study is to assess the genetic diversity in yield and fibre quality traits in 53 cotton genotypes and to analyze the associations among them.

Materials and Methods

Plant Material and Site Characteristics: 50 cotton genotypes along with three checks (NDLH 2051-1, Sivanandhi and CICR 23 Bt) varieties were evaluated during kharif season 2024-2025 at experimental area of Regional Agricultural Research Station, Lam farm, Guntur, Andhra Pradesh which was present at 160 20' North latitude and 800 26' East longitude and an altitude of 31.5 m above mean sea level. The experiment was sown on 1 st August to evaluate genotypes for yield and fibre parameters. The list of the genotypes was presented in (Table 1).

Observations Recorded: Traits like days to fifty percent flowering (plot basis), plant height (cm), number of monopodia per plant, number of sympodia per plant, number of bolls per plant, boll weight (g), lint index (g), seed index (g), ginning out turn (%), seed cotton yield per plant (g) and for fibre quality traits like Upper Half Mean Length (UHML) (mm), uniformity index (%), micronaire (µg/inch) and tenacity (g/tex) data was recorded on plot basis.

Experimental Design: The experiment was conducted using Augmented Block Design (Federer, 1956) [4] in six blocks. In first five blocks, 12 entries (nine genotypes + three checks) and in the last block eight entries (five genotypes + three checks) were planted randomly in the blocks. Each entry was sown in two rows with a row length of 6.3 m each with a spacing of 105 x 60 cm.

Statistical Analysis:

The means of 14 traits were subjected to Analysis of Variance (ANOVA) following the statistical procedure outlined by Falconer (1964) [2] . Subsequently, the mean data was used to perform Principal Component Analysis (PCA) using 'factoextra', 'factoextra' package in R studio version R.4.4.2. In the first Principal Component (PC 1 ), number of bolls per plant made the largest contribution to total divergence (14.82%), followed by days to 50% flowering (12.19%), seed cotton yield per plant (11.63%), upper half mean length (10.15%), uniformity index (8.48%), lint index (8.29%) and number of monopodia per plant (7.97%)). Traits with minimal influence on PC1 included number of sympodia per plant (0.02%), ginning out turn (1.29%), and plant height (1.99%) (Table 2).

In the second Principal Component (PC 2 ), the trait seed index had the strongest impact (15.36%), followed by upper half mean length (12.19%), micronaire (11.78%), lint index (9.91%), seed cotton yield per plant (9.66%), tenacity (8.80%), boll weight (7.38%) and uniformity index (7.22%).

The lowest contributors to PC 2 were number of monopodia per plant (0.27%) and ginning out turn (0.32%) (Table 2).

In third Principal Component (PC 3 ) the trait ginning out turn contributed the highest proportion to the total divergence with a percentage of 28%, followed by number of monopodia per plant, boll weight and number of bolls per plant with the proportions of 17.15%, 13.44% and 10.37% respectively. The traits, number of sympodia per plant (49.41%) and plant height (28.75%) contributed the highest proportion to the total divergence to the PC 4 respectively (Table 2). The highest contribution for the PC 1 was observed by the genotype GP 32 (Suvin) with the proportion of 26.87% followed by the genotypes GP 33 (CCB 29), GP 24 (L 2381) and GP 25 (L 2267) with the percentage of 19.94%, 9.70% and 6.32% respectively. The lowest were contributed by the genotypes GP 31 (NDLA 3116-4), GP 9 (LHBT 6) and GP 5 (LHBT 9) respectively (Table 3). 3).

Based on the biplot analysis of the traits and individual genotypes, the genotypes GP 49 (L 2387), GP 34 (L 2396), GP 40 (L 2281) and GP 39 (L 2278) exhibited better seed index, lint index and seed cotton yield per pant and lowest plant height and these genotypes could be used for improving these traits. The genotypes GP 32 (Suvin) and GP 33 (CCB 29) showed better fibre quality traits like UHML, uniformity index, tenacity and days to fifty percent flowering coupled with lower micronaire value which was in desirable direction. These genotypes could be selected for improving the fibre quality traits. The genotypes GP 24 (L 2381), GP 25 (L 2267), GP 29 (Yaganti), GP 30 (Srinandi) and GP 31 (NDLA 3116-4) reported high micronaire value meant that the genotypes could be considered for coarse fibre genotypes (Figure 1). The analysis thus identified that the yield contributing variables i.e., number of sympodia per plant number of bolls per plant, boll weight, lint index and seed cotton yield per plant were significantly loaded in PC 1 and contributed more towards variability. The PC 2 contributed for fibre quality traits like UHML, tenacity and micronaire (Figure 2). The PCA biplot analysis of 53 cotton genotypes revealed that GP 49 (L2387), GP 34 (L 2396), GP 24 (L 2381), GP 25 (L 2267), GP 29 (Yaganti), GP 30 (Srinandi), GP 31 (NDLA 3116-4), GP 32 (Suvin), GP 33 (CCB 29), GP 37 (L 2275) and GP 45 (L 2385) were identified as the most divergent genotypes as they were located far away from the origin. Hence, these genotypes could be effectively used in hybridization programmes for crop improvement (Figure 1). Similar findings were reported by Shakeel et al. (2015) [14] , Farooq et al. (2017) [3] , Sarwar et al. (2021) [13] , Sahar et al.

(2021) [12] and Chapara et al. (2022) [1] .

Conclusions

Diverse genotypes could be a potential source for transfer of desirable genetic combination for yield improvement. Principal Component Analysis (PCA) identified four principal components with eigen value more than one and attributed 70.56% of cumulative variance. The first two PC' s were considered, as it was capable to explain 49.73% of the total variation. The trait number of bolls per plant contributed the most to total divergence with the proportion of 14.82%, followed by days to fifty percent flowering, seed cotton yield per plant and upper half mean length with the proportions of 12.19%, 11.63% and 10.15%, respectively, in PC 1 . In second Principal Component (PC 2 ), seed index contributed the highest proportion to the total divergence with a percentage of 15.36%, followed by upper half mean length, micronaire, lint index and seed cotton yield per plant with the proportions of 12.19%, 11.78%, 9.90% and 9.66%, respectively. The biplot analysis revealed that GP 49 (L2387), GP 34 (L 2396), GP 24 (L 2381), GP 25 (L 2267), GP 29 (Yaganti), GP 30 (Srinandi), GP 31 (NDLA 3116-4), GP 32 (Suvin), GP 33 (CCB 29), GP 37 (L 2275) and GP 45 (L 2385) were identified as the most divergent genotypes as they were located far away from the origin. Hence, these genotypes could be effectively used in hybridization programmes for crop improvement.

References

1. Chapara R, Reddy KV, Rani MS, Lakshmi BS, Roja V, Pranaya J. Variability studies and genetic divergence in cotton (Gossypium hirsutum L.) germplasm using multivariate analysis. Electronic Journal of Plant Breeding. 2022;13(4):1305-1311.
2. Falconer DS. Introduction to quantitative genetics. New Delhi: Pearson Education India; 1964.
3. Farooq J, Rizwan M, Sharif I, Saleem S, Chohan SM, Kainth RA. Genetic diversity studies in some advanced lines of Gossypium hirsutum L. for yield and quality related attributes using cluster and principal component analysis. AA Bioflux. 2017;9(3):111-118.
4. Federer WT. Augmented designs, Hawaiian Planters' Record. 1956;55:191-208.
5. ICAR-All India Coordinated Research Project on Annual Report 2024-25. Nagpur: ICAR; 2025.
6. Isong A, Balu PA, Ramakrishnan P. Association and principal component analysis of yield and its components in cultivated cotton. Electronic Journal of Plant Breeding. 2017;8(3):857-864.
7. Komala M, Ganesan NM, Kumar M. Genetic variability, heritability and correlation analysis in F₂ populations of ratoon upland cotton hybrids. International Journal of Agriculture, Environment and Biotechnology. 2018;11(6):815-827.
8. Ruilian L, Zheng L, Jinxiang C. Comparison of fibre quality characteristics of different varieties of hybrid cotton. Journal of Hunan Agricultural University. 2005;31(5):492-495.
9. McCarty Jenkins JN, Wu J. Primitive accession derived germplasm by cultivar crosses as sources for cotton improvement: I. Phenotypic values and variance components. Crop Science. 2004;44(4):1226-1230.
10. Naik BM, Satish Y, Babu DR. Genetic diversity analysis in American cotton (Gossypium hirsutum L.). Electronic Journal of Plant Breeding. 2016;7(4):10021006.
11. Percival AE, Kohel RJ. Distribution, collection, and evaluation of Gossypium. Advances in Agronomy. 1990;44:225-256.