Next Article in Journal
Genetic Diversity of Lowbush Blueberry throughout the United States in Managed and Non-Managed Populations
Next Article in Special Issue
Multi-Trait Diverse Germplasm Sources from Mini Core Collection for Sorghum Improvement
Previous Article in Journal
The Effects of Individual Variables, Farming System Characteristics and Perceived Barriers on Actual Use of Smart Farming Technologies: Evidence from the Piedmont Region, Northwestern Italy
Previous Article in Special Issue
Beyond Bird Feed: Proso Millet for Human Health and Environment
Article Menu
Issue 5 (May) cover image

Export Article

Agriculture 2019, 9(5), 112; https://doi.org/10.3390/agriculture9050112

Article
Variability in the Global Proso Millet (Panicum miliaceum L.) Germplasm Collection Conserved at the ICRISAT Genebank
Genebank, International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru, Hyderabad 502324, India
*
Author to whom correspondence should be addressed.
Received: 23 March 2019 / Accepted: 6 May 2019 / Published: 24 May 2019

Abstract

:
Proso millet (Panicum miliaceum L.), also known as common millet or broomcorn millet, is an important ancient crop mostly grown for food, feed, and fodder purposes largely in China, Russia, India, and the USA. It is an under-researched and under-utilized crop. Over 29,000 germplasm accessions have been conserved in genebanks globally. Five races (miliaceum, patentissimum, contractum, compactum, ovatum) have been recognized in proso millet based on panicle morphology and shape. The genebank at the International Crops Research Institute for the Semi-Arid Tropics conserves 849 accessions of proso millet originating from 30 countries and represents all five races. Characterization of these germplasm accessions revealed large variability for morpho-agronomic traits, including for days to 50% flowering (26 to 50 days), plant height (20 to 133 cm), and inflorescence length (22 to 400 mm). On average, the race miliaceum was tall (62 cm) with long panicles (209 mm) and ovatum had short plants (46 cm) with small panicles (108 mm). The average Gower’s distance based on 18 morpho-agronomic traits on 841 accessions was 0.261. The race miliaceum had the highest among accessions within race average pairwise distance (0.254), while the distance was the lowest in ovatum (0.192). The races miliaceum and ovatum showed the highest divergence with each other (0.275), while the lowest divergence was observed between compactum and ovatum (0.229). Trait-specific sources were identified for early maturity, tall plants, long inflorescences, and greater seed size. The information on variability and trait-specific sources identified could potentially support proso millet improvement.
Keywords:
proso millet; germplasm; diversity; race; trait-specific sources

1. Introduction

Proso millet (Panicum miliaceum L.), also called as common millet or broomcorn millet, is an annual herbaceous plant from the genus Panicum, and it has a chromosome number of 2n = 36 with a basic chromosome number of x = 9. de Wet [1] classified proso millet germplasm into five races (miliaceum, patentissimum, contractum, compactum, ovatum) based on panicle morphology and shape (Figure S1). The race miliaceum resembles wild P. miliaceum in inflorescence morphology, characterized by large open inflorescences with suberect branches that are sparingly subdivided. The race patentissimum is characterized by slender and diffuse panicle branches, which is often difficult to distinguish from race miliaceum. Accessions with more or less compact inflorescences are classified into races contractum, compactum, and ovatum. Accessions in the race contractum have compact drooping inflorescences, the race compactum has cylindrical inflorescences that are essentially erect, while accessions with compact and slightly curved inflorescences that are ovate in shape belong to race ovatum [1]. Vavilov [2] suggested China as the center of diversity for proso millet, while Harlan [3] opined that proso millet was probably domesticated in China and Europe. The earliest records come from the Yellow River valley site of Cishan, China dated between 10,300 and 8700 cal Before Present (BP) [4]. Evidence of proso millet also occurs at a number of pre-7000 cal BP sites in Eastern Europe, in the form of charred grains and grain impressions in pottery [5]. These two centers of earlier records suggest the independent domestication of proso millet in Eastern Europe or Central Asia, or may have also originated from a domestication within China and then spread westward across the Eurasian steppe [6].
Proso millet is grown in Asia, Australia, North America, Europe, and Africa, and is used for feeding birds and as livestock feed in developed countries and for food in some parts of Asia [7]. Proso millet is cultivated in about 0.82 million ha in Russia, 0.32 million ha in China [8], 0.20 million ha in the U.S [9], 0.03 million ha in India [10], and 0.002 million ha in Korea [11]. The U.S. is among the top producers, and exports 15–20% of its annual proso millet production to over 70 countries, primarily as feed [9]. This crop matures in 6 to 12 weeks and requires less water and adapts well to varied environmental conditions [12]. Proso millet grains are rich in protein, vitamins, minerals, and micronutrients including iron, zinc, copper, and manganese, compared to other staple cereals [13]. The protein content of proso millet is comparable to that of wheat, and its grains are richer in essential amino acids (leucine, isoleucine, and methionine) than those of wheat [14,15]. The husked grains are eaten whole, boiled, or cooked like rice (Oryza sativa L.), and are sometimes ground to make roti (flatbread). The starch is suitable as a sizing agent in the textile industry. Green plants are used as fodder for cattle and horses, and are also used as hay. Proso millet is used to make fermented beverages in Africa and Asia, and is receiving growing interest from food industries in Europe and North America because of its mild flavor, light color, gluten-free quality, and potential health benefits [16].
Proso millet is an under-researched and under-utilized crop. Germplasm plays an important role in crop improvement. Considerable numbers of proso millet germplasms have been conserved in genebanks globally. Information about variability in the germplasm collection of a given species for important traits, including yield and quality, enables their utilization in crop improvement programs. The genebank at the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) conserves 849 accessions of proso millet originating from 30 countries. This study aims to assess the racial characteristics and geographical distribution, as well as the diversity of the proso millet germplasm collection conserved at the ICRISAT genebank for various morpho-agronomic traits.

2. Materials and Methods

2.1. Experiment Details

The genebank at ICRISAT, Patancheru conserves 849 proso millet germplasm accessions originating from 30 countries. They were characterized from 1977 onwards at ICRISAT, Patancheru located at (17°53′ N, 78°27′ E, 545 m asl), as and when new germplasm entries were received at the ICRISAT genebank. Sowings were done during second fortnight of July in each year, on red soils (Alfisols), following augmented block design, along with controls. Each accession was planted in a single row of 4 m length, 60 cm between rows, with plant-to-plant spacing of 10 cm and at uniform depth. Diammonium phosphate was applied at 100 kg ha−1 as a basal dose to supply nitrogen and phosphorus. In addition, 100 kg ha−1 of urea was applied as top dressing. The precision fields at the ICRISAT center have uniform fertility and a gentle slope of 0.5%. All cultural practices and data recordings were the same for all years of evaluation. The average annual rainfall at this location is about 890 mm (averaged from 1977 to 2018), which normally occurs during June to September.

2.2. Data Collection

Data on various qualitative traits, namely, growth habit (erect, decumbent, erect geniculate, and prostrate), culm branching (low, medium, and high), sheath pubescence (sparse, medium, and dense), ligule pubescence (sparse, medium, and dense), leaf pubescence (sparse, medium, and dense), inflorescence shape (arched sparse, arched dense, diffuse sparse, diffuse dense, elliptic sparse, elliptic dense, globose sparse, and globose dense), seed color (white, light red, straw, dark red, dark green, olive green, dark olive green, light brown, brown, dark brown, black), and apiculus color (straw and purple); and quantitative traits, namely, days to 50% flowering, plant height (cm), basal tillers number, flag leaf blade length (mm), flag leaf blade width (mm), flag leaf sheath length (mm), peduncle length (mm), panicle exsertion (mm), inflorescence length (mm), and inflorescence primary branches number were recorded following descriptors for Panicum miliaceum [17].

2.3. Data Analyses

Data collected in the individual years were analyzed independently and stored in the database (also available in the ICRISAT genebank webpage http://genebank.icrisat.org/) and used in this study. The range and means were calculated for all traits, for each race, and for each country of origin. The mean performances of races for each quantitative trait were compared using Newman–Keuls test [18,19], and the homogeneity of variances was tested using Levene’s procedure [20] using the R packages “agricolae” [21] and “car” [22]. Principal component analysis (PCA) based on 10 quantitative traits was performed to determine the relative importance of different traits in capturing the variation in proso millet collection, and the Shannon–Weaver diversity index (H′) [23] was used as a measure of the phenotypic diversity of eight qualitative and ten quantitative traits, using GenStat 17th Edition (https://www.vsni.co.uk/). Gower’s dissimilarity matrix [24] was constructed using both qualitative and quantitative traits using the R package “cluster” [25]. The Gower’s dissimilarity matrix was then used to identify most diverse pairs of accessions, and accessions were clustered following the neighbor-joining method [26] using the software DARwin 6.0.14 [27].

3. Results

3.1. Racial and Geographical Distribution

Accessions of proso millet can be classified into five races: miliaceum, patentissimum, contractum, compactum, and ovatum, based on panicle morphology and shape [1]. The 849 accessions of proso millet conserved in the ICRISAT genebank were classified into five races, namely, miliaceum (63.5%), compactum (11.5%), contractum (10.8%), ovatum (5.7%), and patentissimum (7.4%) (Table 1). The majority of proso millet accessions conserved in the ICRISAT genebank originated from Asia (37.1%) and Europe (18.5%). At the country level, 14% of the entire collection was from the Russian Federation, while about 9% each were from India and Republic of Korea, 4–6% were from Syria (4.1%), Pakistan (4.8%), and Turkey (5.8%), and the remaining countries represented <2% (Table 1).

3.2. Racial and Geographical Diversity

3.2.1. Qualitative Traits

The frequency distributions of different phenotypic classes of the eight qualitative traits showed considerable variation (Table 2). In the entire set, decumbent growth habit (74.6%), high culm branching (45.0%), medium sheath pubescence (40.6%), sparse and medium ligule pubescence (45.5% and 39.9%, respectively), sparse leaf pubescence (51.1%), diffuse sparse and diffuse dense inflorescence shape (30.2% and 27.7%, respectively), light brown-colored seed (42.1%), and straw apiculus color (63.2%) were the predominant classes of qualitative traits. Among the races, the frequencies of qualitative trait classes varied, particularly for culm branching, sheath pubescence, and inflorescence shape. High culm branching was the most common trait in all races except patentissimum, where low (41.3%) and medium (34.9%) branching were the more prevalent, while contractum had all the three classes of culm branching—medium (39.6%), low (34.1%), and high (27.5%), in high frequency. Accessions of the race compactum and miliaceum had all three classes of sheath pubescence in high frequency (26.5–38.8% in compactum; 26.3–42.7% in miliaceum), while contractum had medium sheath pubescence (46.2%), ovatum showed sparse (43.8%) sheath pubescence, and patentissimum showed dense (39.7%) and sparse (36.5%) pubescence. For inflorescence shape, arched dense inflorescence in contractum (80.2%), elliptic dense (40.9%) and elliptic sparse (48.0%) inflorescence in compactum, diffuse sparse (43.4%) and diffuse dense (41.2%) inflorescence in miliaceum, globose dense (54.2%) and globose sparse (35.4%) inflorescence in ovatum, and arched sparse (68.3%) inflorescence in patentissimum were in high proportion. Light-brown seed was in higher proportion in all five races, except in ovatum where both straw (45.8%) and light-brown (27.1%) colored seeds were more prevalent. Race miliaceum had all eleven classes of seed color (Table 2). Regarding other traits, a large portion of accessions in all five races had decumbent growth habit (58.2–89.6%), sparse ligule pubescence (40.6–58.3%), and sparse leaf pubescence (45.8–64.6%), and straw apiculus color (53.8–89.6%). The H′ revealed that in the entire set, inflorescence shape had the highest diversity (0.767) while seed color had the highest H′ in each race (0.582 in ovatum to 0.709 in compactum). Figure 1 shows variability in the proso millet germplasm accessions for seed color. The lowest H′ was observed for apiculus color in the entire set (0.286) (Table 3).

3.2.2. Quantitative Traits

Considerable variability was observed in the entire collection of proso millet germplasm conserved at the ICRISAT genebank: days to 50% flowering varied from 26 to 50 days after sowing (DAS), plant height varied from 20 to 133 cm and inflorescence length varied from 22 to 400 mm. Figure S2 shows frequency (histogram) of all the ten quantitative traits. On average, accessions of the race ovatum flowered at 33 DAS and produced short plants (46 cm) and short inflorescence (108 mm), whereas miliaceum flowered at 35 DAS and produced tall plants (62 cm) and long inflorescences (209 mm) with high number of inflorescence primary branches (17); however, a wide range of variation existed within each race (Table 4). At the country level, early flowering and shortest inflorescence accessions were from Russian Federation (average 31 DAS), and late flowering were from India (average 39 DAS), the shortest plants (average 34 cm) were from Mexico, and the tallest plants (average 93 cm) and longest inflorescences (average 307 mm) were from Nepal (data not shown). The H′ revealed that the trait inflorescence primary branch number had the highest diversity (0.649) in the entire set as well as in the races miliaceum (0.649) and patentissimum (0.586), and in Asia (0.631) and Europe (0.651), while basal tillers number in compactum (0.643), flag leaf blade length in contractum (0.624), plant height in ovatum (0.625), and days to 50% flowering (0.544) and flag leaf sheath length (0.544) in the Americas had the highest H′ value. However, all the quantitative traits showed high diversity in the entire set (0.576–0.649) as well as in each race (0.315–0.649) and region of origin (0.276–0.651) (Table 3).
Gower’s phenotypic distance constructed based on 18 traits including 8 qualitative traits and 10 quantitative traits using 841 accessions revealed an average distance of 0.261, varying from 0.010 between IPm 2011 and IPm 2012 to 0.591 between IPm 370 and IPm 2806, and the top ten pairs of the most diverse accessions were identified (Table 5). On average, miliaceum had the highest among accessions within race pairwise distance (0.254) while the lowest was seen in ovatum (0.192). The races miliaceum and ovatum showed the highest divergence with each other (0.275), while the lowest divergence was observed between compactum and ovatum (0.229) (Table 6). Clustering of accessions based on Gower’s distance matrix following the neighbor-joining method revealed three major groups of accessions: Cluster I (C-I) largely contained accessions from Asia, Cluster II (C-II) represented accessions from Europe, and Cluster III (C-III) represented accessions from both Asia and Europe (Figure 2). Accessions of miliaceum, contractum, and patentissimum were found in all three clusters, while those of ovatum and contractum were found mostly in C-II and C-III. Principal component analysis revealed the first three principal components (PCs) as important, explaining about 73.32% of the total variance. Plant height, flag leaf blade length, flag leaf sheath length, and inflorescence length contributed largely to PC1, which explained about 37.19% of the total variance (Table 7).

3.2.3. Trait-Specific Sources

In the entire collection, 12 accessions flowered in <30 days (26 to 29 DAS), eight of them from the Russian Federation (IPm 2577, IPm 2601, IPm 2526, IPm 2510, IPm 2527, IPm 2774, IPm 2509, IPm 2521), one from Syria (IPm 2903), and three with unknown origin (IPm 2273, IPm 2035, IPm 2007); six accessions had inflorescence length of over 350 mm (IPm 2661, IPm 2660, IPm 2095, IPm 2198, IPm 2107, IPm 2197); and four accessions (IPm 2198, IPm 2100, IPm 2197, IPm 2200) were over 125 cm tall (125–133 cm); 8 accessions (IPm 2535, IPm 370, IPm 381, IPm 79, IPm 121, IPm 2140, IPm 2122, IPm 38) had a maximum seed length of 3.5–4.3 mm.

3.2.4. Trait Associations

Correlations among agronomic traits were estimated in the entire set and in each race (Table 8). In the entire set, days to 50% flowering was significantly and positively correlated with plant height, basal tillers number, flag leaf blade length, and inflorescence length, while it was significantly negatively correlated with flag leaf blade width, peduncle length, and panicle exsertion. Peduncle length and panicle exsertion in the entire set and in each race showed significant negative correlations with days to 50% flowering (except panicle exsertion in ovatum, which showed positive correlation with days to 50% flowering). Days to 50% flowering in the entire set and in miliaceum showed significantly positive correlation with plant height, while ovatum showed negatively significant correlation. Similarly, inflorescence length in the entire set showed significant positive correlation with days to 50% flowering, plant height, basal tillers number, flag leaf blade length, flag leaf blade width, flag leaf sheath length, peduncle length, and inflorescence primary branches number; however, the magnitude of correlations differed within each race. For instance, inflorescence length showed significant positive correlations: with peduncle length and panicle exsertion in ovatum; with days to 50% flowering, plant height, flag leaf blade length, flag leaf sheath length, and inflorescence primary branches number in patentissimum; with all traits in miliaceum; and with plant height, basal tillers, flag leaf blade length, flag leaf blade width, and flag leaf sheath length in contractum and compactum. This signifies that while estimating correlation, we must also consider existing racial characteristics/groups because correlations depend on the nature of the population.

4. Discussion

Germplasm without sufficient characterization and evaluation data—particularly in case of low-research-priority crops like proso millet, limiting the use of germplasm in breeding programs due to extremely low funding for research and development compared to other major crops. This study provides a broad overview of variability in the global proso millet germplasm conserved at the ICRISAT genebank for morpho-agronomic traits, and investigated racial and geographical diversity. The majority of accessions in the ICRISAT proso millet collection belonged to miliaceum (63.5%), while the other races represented <11%, indicating the low representation of other races in the entire set. de Wet [1] classified proso millet germplasm into five races based on panicle morphology and shape. Besides these characteristics, accessions also showed variations in different morpho-agronomic traits among the races. Frequencies of qualitative traits varied among the races, particularly for culm branching, sheath pubescence, and inflorescence shape. Most accessions in the ovatum flowered earlier, produced short plants, and short inflorescence, while miliaceum flowered late, produced tall plants, and long inflorescence. Accessions within miliaceum also had the highest phenotypic distance (0.254), while the lowest distance was seen in ovatum (0.192). Races miliaceum and ovatum highly diverged with each other (0.275), while low divergence was observed between compactum and ovatum (0.229).
In the ICRISAT genebank proso millet collection, the majority of accessions originated from Asia (37.1%) and Europe (18.5%), indicating these two regions as major centers of diversity for proso millet. The highest diversity (H′) was found in accessions from Asia (0.497) compared to those from Europe (0.431). Neighbor-joining clustering of accessions revealed three major clusters, representing those from Asia in C-I and Europe in C-II, with C-III representing those from both regions. The results from this study support the independent domestication of proso millet in Central Asia and Eastern Europe, or indicate that they might also have originated from a domestication within China and then spread westward across the Eurasian steppe [5]. Early flowering and shortest inflorescence accessions were largely from the Russian Federation (average 31 DAS), and late flowering accessions were from India (average 39 DAS), the shortest plants (average 34 cm) were from Mexico, and the tallest plants (average 93 cm) and longest inflorescences (average 307 mm) were from Nepal. These sets of accessions could be selectively explored for the identification of useful germplasms for respective traits.
The trait-specific sources identified in this study for early maturity, tall plants, long inflorescences, and greater seed size from the entire set of proso millet germplasms conserved at the ICRISAT genebank, and those that were identified in our previous study [12] were as follows: 18 accessions for high grain yield (IPm 9, IPm 2784, IPm 2621, IPm 2802, IPm 390, IPm 361, IPm 2824, IPm 2783, IPm 366, IPm 2620, IPm 2660, IPm 2685, IPm 2158, IPm 388, IPm 2700, IPm 384, IPm 1545, IPm 2661; grain yield 1601–2334 kg ha−1), 8 accessions with greater seed weight (IPm 362, IPm 2826, IPm 381, IPm 2575, IPm 2273, IPm 2769, IPm 2780, IPm 2037; 100-seed weight 0.60–0.66g); 2 accessions that produced high grain yield with greater seed size (IPm 2, IPm 2661); 12 accessions for high grain Fe (63.3–73.2 mg kg−1); 27 for Zn (40.6–46.7 mg kg−1), 56 for Ca (185.5–241.2 mg kg−1), and 27 for protein (16%–19%) including IPm 2069, IPm 2076, and IPm 2537 rich in grain Fe, Zn, Ca, and protein contents [12] could potentially support the breeding of high-yielding nutrient-dense cultivars with broad genetic base in proso millet. Besides these sources, the top ten pairs of the most diverse accessions were identified that could be utilized to broaden the genetic base of proso millet cultivars.
The ICRISAT genebank supplied over 6900 seed samples of proso millet to researchers in 39 countries for use in proso millet improvement. Proso millet researchers can obtain seed samples from the ICRISAT genebank (http://genebank.icrisat.org/) for research purposes via a Standard Material Transfer Agreement.

Supplementary Materials

The following are available online at https://www.mdpi.com/2077-0472/9/5/112/s1, Figure S1: Five races of proso millet (miliaceum, patentissimum, contractum, compactum, ovatum) based on panicle morphology and shape, Figure S2: Frequency (histogram) of ten quantitative traits of proso millet germplasm conserved at the ICRISAT genebank, Patancheru, India.

Author Contributions

Conceptualization, M.V., H.D.U., and V.C.R.A.; Performed the experiments: M.V., H.D.U.; Analyzed the data, M.V., D.N.; Contributed to the writing of the manuscript, M.V., H.D.U., and V.C.R.A.

Funding

We gratefully acknowledge support from the CGIAR Genebank Platform coordinated by the Crop Trust for supporting genebank activities.

Acknowledgments

The authors gratefully acknowledge the contributions of Research Technicians and Scientific Officers and database managers of the ICRISAT Genebank, Patancheru, India in the data collection and documentation.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. De Wet, J.M.J. Origin, evolution and systematics of minor cereals. In Small Millets in Global Agriculture; Seetharam, A., Riley, K.W., Harinarayana, G., Eds.; Oxford & IBH Publishing Co. Pvt. Ltd.: New Delhi, India, 1986; pp. 19–30. [Google Scholar]
  2. Vavilov, N.I. Centers of origin of cultivated plants. Tr po Prikl Bot Genet Sel [Bull Appl Bot & Genet Sel] 1926, 16, 139–248. (In Russian) [Google Scholar]
  3. Harlan, J.R. Crops and Man. American Society of Agronomy; Crop Science Society of America, Inc.: Madison, WI, USA, 1975. [Google Scholar]
  4. Lu, H.; Zhang, J.; Liu, K.-B.; Wu, N.; Li, Y.; Zhou, K.; Ye, M.; Zhang, T.; Zhang, H.; Yang, X.; et al. Earliest domestication of common millet (Panicum miliaceum) in East Asia extended to 10,000 years ago. Proc. Acad. Sci. 2009, 106, 7367–7372. [Google Scholar] [CrossRef] [PubMed]
  5. Hunt, H.V.; Linden, M.V.; Liu, X.; Motuzaite-Matuzeviciute, G.; Colledge, S.; Jones, M.K. Millets across Eurasia: chronology and context of early records of the genera Panicum and Setaria from archaeological sites in the Old World. Veg. Hist. Archaeobotany 2008, 17, 5–18. [Google Scholar] [CrossRef] [PubMed]
  6. Hunt, H.V.; Campana, M.G.; Lawes, M.C.; Park, Y.; Bower, M.A.; Howe, C.J.; Jones, M.K. Genetic diversity and phylogeography of broomcorn millet (Panicum miliaceum L.) across Eurasia. Mol. Ecol. 2011, 20, 4756–4771. [Google Scholar] [CrossRef] [PubMed]
  7. Rajput, S.G.; Plyler-Harveson, T.; Santra, D.K. Development and Characterization of SSR Markers in Proso Millet Based on Switchgrass Genomics. Am. J. Plant Sci. 2014, 5, 175–186. [Google Scholar] [CrossRef]
  8. Diao, X. Production and genetic improvement of minor cereals in China. Crop. J. 2017, 5, 103–114. [Google Scholar] [CrossRef]
  9. Habiyaremye, C.; Matanguihan, J.B.; Guedes, J.D.; Ganjyal, G.M.; Whiteman, M.R.; Kidwell, K.K.; Murphy, K.M. Proso Millet (Panicum miliaceum L.) and Its Potential for Cultivation in the Pacific Northwest, U.S.: A Review. Front. Plant Sci. 2017, 7, 501. [Google Scholar] [CrossRef] [PubMed]
  10. Bhat, B.V.; Tonapi, V.A.; Rao, B.D.; Singode, A.; Santra, D. Production and utilization of millets in India. In Proceedings of the International Millet Symposium and the 3rd International Symposium on Broomcorn Millet (3rd ISBM), Fort Collins, CO, USA, 8–12 August 2018; pp. 24–26. [Google Scholar]
  11. Park, C.H. Production and Utilization of Broomcorn Millet in Korea. In Proceedings of the International Millet Symposium and the 3rd International Symposium on Broomcorn Millet (3rd ISBM) Program and Abstracts, Fort Collins, CO, USA, 8–12 August 2018; p. 27. [Google Scholar]
  12. Vetriventhan, M.; Upadhyaya, H.D. Diversity and trait-specific sources for productivity and nutritional traits in the global proso millet (Panicum miliaceum L.) germplasm collection. Crop. J. 2018, 6, 451–463. [Google Scholar] [CrossRef]
  13. Gomeshe, S.S. Proso millet, Panicum miliaceum (L.): Genetic improvement and research needs. In Millets and Sorghum: Biology and Genetic Improvement; Patil, J.V., Ed.; John Wiley & Sons Ltd.: Chichester, UK, 2017; pp. 150–179. [Google Scholar]
  14. Kalinová, J.; Moudrý, J. Content and Quality of Protein in Proso Millet (Panicum miliaceum L.) Varieties. Plant Foods Hum. Nutr. 2006, 61, 43–47. [Google Scholar] [CrossRef] [PubMed]
  15. Saleh, A.S.; Zhang, Q.; Chen, J.; Shen, Q. Millet Grains: Nutritional Quality, Processing, and Potential Health Benefits. Compr. Rev. Food Sci. Food Saf. 2013, 12, 281–295. [Google Scholar] [CrossRef]
  16. Wang, R.; Hunt, H.V.; Qiao, Z.; Wang, L.; Han, Y. Diversity and Cultivation of Broomcorn Millet (Panicum miliaceum L.) in China: A Review. Econ. Bot. 2016, 70, 332–342. [Google Scholar] [CrossRef]
  17. IBPGR. Descriptors for Panicum miliaceum and P. sumatrense; IBPGR: Rome, Italy, 1985. [Google Scholar]
  18. Newman, D. The distribution of range in samples from a normal population, expressed in terms of an independent estimate of standard deviation. Biometrika 1939, 31, 20–30. [Google Scholar] [CrossRef]
  19. Keuls, M. The use of the “Studentized range” in connection with an analysis of variance. Euphytica 1952, 1, 112–122. [Google Scholar] [CrossRef]
  20. Levene, H. Robust Tests for Equality of Variances. In Contributions to Probability and Statistics: Essays in Honor of Harold Hotelling; Olkin, I., Ghurye, S.G., Hoeffding, W., Madow, W.G., Mann, H.B., Eds.; Stanford University Press: Palo Alto, CA, USA, 1960; pp. 278–292. [Google Scholar]
  21. de Mendiburu, F. agricolae: Statistical Procedures for Agricultural Research. R package version 1.2-7. Comprehensive R Arch. Network. Available online: https://cran.r-project.org/package=agricolae (accessed on 21 May 2018).
  22. Fox, J.; Weisberg, S. An {R} Companion to Applied Regression, 2nd ed.; Sage: Thousand Oaks, CA, USA, 2011; Available online: http://socserv.socsci.mcmaster.ca/jfox/Books/Companion (accessed on 25 October 2018).
  23. Shannon, C.E.; Weaver, W. The Mathematical Theory of Communication; The University of Illinois Press: Urbana, IL, USA, 1949. [Google Scholar]
  24. Gower, J.C. A general coefficient of similarity and some of its properties. Biometrics 1971, 27, 857–874. [Google Scholar] [CrossRef]
  25. Maechler, M.; Rousseeuw, P.; Struyf, A.; Hubert, M.; Hornik, K. Cluster: Cluster Analysis Basics and Extensions, R Package Version 2.0.7-1. 2017.
  26. Saitou, N.; Nei, M. The neighbour-joining method: A new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 1987, 4, 406–425. [Google Scholar] [PubMed]
  27. Perrier, X.; Jacquemoud-Collet, J.P. DARwin Software. 2006. Available online: http://darwin.cirad.fr/darwin (accessed on 3 August 2018).
Figure 1. Variability in the proso millet germplasm accessions for seed color.
Figure 1. Variability in the proso millet germplasm accessions for seed color.
Agriculture 09 00112 g001
Figure 2. Neighbor-joining clustering of proso millet accessions based on morpho-agronomic traits. Color code: pink—Europe; green—Asia; blue—the Americas.
Figure 2. Neighbor-joining clustering of proso millet accessions based on morpho-agronomic traits. Color code: pink—Europe; green—Asia; blue—the Americas.
Agriculture 09 00112 g002
Table 1. Proso millet germplasm conserved at the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) genebank, Patancheru, India.
Table 1. Proso millet germplasm conserved at the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) genebank, Patancheru, India.
CountryRaceUnclassifiedTotal
compactumcontractummiliaceumovatumpatentissimum
Russian Federation281364511 121
India 357 8876
Republic of Korea21652 3 73
Turkey863032 49
Pakistan1131 8 41
Syria1101724135
Afghanistan5 11 16
Mexico 4324 13
Germany129 12
Hungary116 2 10
Iran1 8 9
Nepal 4 2 6
Ukraine1 3 4
United Kingdom 12 1 4
Australia 2 2
Bangladesh 2 2
China1 1 2
Iraq2 2
Kazakhstan 1 1 2
Sri Lanka 2 2
Argentina 1 1
Canada 1 1
Japan1 1
Kenya 1 1
Kyrgyzstan 1 1
Lebanon 1 1
Malawi 1 1
Romania 1 1
Spain 1 1
Socialist Federal Republic of Yugoslavia 1 1
Unknown origin45292343515 358
Total989253948639849
Table 2. Frequency of phenotypic classes of each qualitative trait in the five races and in the entire set of proso millet germplasm conserved at the ICRISAT genebank, Patancheru, India.
Table 2. Frequency of phenotypic classes of each qualitative trait in the five races and in the entire set of proso millet germplasm conserved at the ICRISAT genebank, Patancheru, India.
TraitClassRaceEntire Set
compactumcontractummiliaceumovatumpatentissimum
Growth habitDecumbent85 (86.7) †53 (58.2)407 (75.5)43 (89.6)38 (60.3)627 (74.6)
Erect11 (11.2)24 (26.4)93 (17.3)3 (6.3)18 (28.6)149 (17.8)
Erect geniculate2 (2.0)14 (15.4)39 (7.2)2 (4.2)7 (11.1)64 (7.6)
Culm branchingHigh48 (49.0)25 (27.5)269 (49.9)21 (43.8)15 (23.8)379 (45.0)
Low14 (14.3)31 (34.1)93 (17.3)1 (2.1)26 (41.3)165 (19.6)
Medium36 (36.7)36 (39.6)177 (32.8)26 (54.2)22 (34.9)297 (35.4)
Sheath pubescenceDense26 (26.5)23 (25.3)167 (31.0)7 (14.6)25 (39.7)249 (29.5)
Medium34 (34.7)42 (46.2)230 (42.7)20 (21.7)15 (23.8)341 (40.6)
Sparse38 (38.8)27 (29.7)142 (26.3)21 (43.8)23 (36.5)251 (29.9)
Ligule pubescenceDense17 (17.3)8 (9.9)85 (15.8)4 (8.3)9 (14.3)124 (14.6)
Medium30 (30.6)34 (37.4)235 (43.6)16 (33.3)20 (31.7)335 (39.9)
Sparse51 (52.0)50 (54.9)219 (40.6)28 (58.3)34 (54.0)382 (45.5)
Leaf pubescenceDense17 (17.3)8 (8.8)79 (14.7)4 (8.3)6 (9.5)114 (13.6)
Medium26 (26.5)28 (30.8)213 (39.5)13 (21.1)17 (27.0)297 (35.4)
Sparse55 (56.1)56 (61.5)247 (45.8)31 (64.6)40 (63.5)430 (51.1)
Inflorescence shapeArched dense3 (3.1)73 (80.2)47 (8.7)-1 (1.6)124 (14.8)
Arched sparse2 (2.0)5 (5.5)13 (2.4)-43 (68.3)63 (7.5)
Diffuse dense1 (1.0)2 (2.2)222 (41.2)-8 (12.7)233 (27.7)
Diffuse sparse4 (4.1)3 (3.3)234 (43.4)2 (4.2)11 (17.5)254 (30.2)
Elliptic dense40 (40.9)2 (2.2)16 (3.0)1 (2.1)-59 (7.0)
Elliptic sparse47 (48.0)7 (7.7)6 (1.1)2 (4.2)-62 (7.4)
Globose dense1 (1.0)--26 (54.2)-27 (3.2)
Globose sparse--1 (0.2)17 (35.4)-18 (2.1)
Seed colorBlack-1 (1.1)2 (0.4)--3 (0.4)
Brown--4 (0.7)--4 (0.5)
Dark brown3 (3.1)1 (1.1)8 (1.5)--12 (1.4)
Dark green-1 (1.1)4 (0.7)--5 (0.6)
Dark olive green1 (1.0)-39 (7.2)-3 (4.8)43 (5.1)
Dark red12 (12.2)2 (2.2)8 (1.5)2 (4.2)-24 (2.9)
Light brown35 (35.7)28 (30.8)255 (47.3)13 (27.1)23 (36.5)354 (42.1)
Light red9 (9.2)7 (7.8)31 (5.8)5 (10.4)2 (3.2)55 (6.5)
Olive green-2 (2.2)13 (2.4)-4 (6.3)19 (2.3)
Straw17 (17.3)28 (30.8)93 (17.3)22 (45.8)20 (31.7)180 (21.4)
White21 (21)22 (24.2)82 (15.2)6 (12.5)11 (17.5)142 (16.9)
Apiculus colorPurple24 (24.5)19 (20.9)249 (46.2)5 (10.4)12 (19.0)309 (36.8)
Straw74 (75.5)73 (80.2)290 (53.8)43 (89.6)51 (81.0)531 (63.2)
values within parentheses indicate the percentage of accessions in each phenotypic classes of each qualitative trait.
Table 3. Shannon–Weaver diversity indices (H′) of eight qualitative and ten quantitative traits in the entire set, five races and regions of proso millet germplasm conserved at the ICRISAT genebank, Patancheru, India.
Table 3. Shannon–Weaver diversity indices (H′) of eight qualitative and ten quantitative traits in the entire set, five races and regions of proso millet germplasm conserved at the ICRISAT genebank, Patancheru, India.
TraitEntireRace Region
compactumcontractummiliaceumovatumpatentissimumAmericasAsiaEurope
Qualitative traits
Growth habit0.3130.1950.4150.3060.1760.3940.1710.4360.293
Culm branching0.4550.4320.4720.4410.3360.4670.3740.4360.460
Sheath pubescence0.4720.4720.4620.4680.4370.4670.4130.4300.320
Ligule pubescence0.4370.4370.3960.4430.3860.4240.3710.4580.293
Leaf pubescence0.4260.4260.3810.4370.3660.3760.1710.4540.272
Inflorescence shape0.7670.4900.3540.5200.4540.3880.6470.7280.732
Seed color0.7160.7090.6850.6980.5820.6370.5080.7370.830
Apiculus color0.2860.2420.2210.3000.1450.2120.1710.2940.252
Mean0.4840.4250.4230.4510.3600.4200.3530.4970.431
Quantitative traits
Days to 50% flowering0.5850.5680.5880.5720.3150.4840.5440.5940.478
Plant height (cm)0.5760.5670.6070.6020.6250.5650.4610.6210.600
Basal tillers number0.5930.6430.6030.6040.5660.5620.2760.5400.579
Flag leaf blade length (mm)0.6180.6290.6240.6000.5680.5190.5410.6220.609
Flag leaf blade width (mm)0.6090.4200.5580.5580.5060.5560.4710.6050.531
Flag leaf sheath length (mm)0.6350.5080.5910.6260.5650.5560.5440.5850.569
Peduncle length (mm)0.6220.6140.5840.6250.6200.5790.5090.6040.592
Panicle exsertion (mm)0.5980.5710.5840.5950.5260.5840.4710.6000.604
Inflorescence length (mm)0.6280.4910.5610.6260.5980.5610.4710.6020.598
Inflorescence primary branches number0.6490.5980.6040.6490.6010.5860.4030.6310.651
Mean0.6110.5610.5900.6060.5490.5550.4690.6000.581
Table 4. Mean and range of proso millet germplasm conserved at the ICRISAT genebank for agronomic traits.
Table 4. Mean and range of proso millet germplasm conserved at the ICRISAT genebank for agronomic traits.
TraitMean $ Range
CP #CTMLOVPAEntire set (849 Accessions)CPCTMLOVPAEntire set (849 Accessions)
DF 33 c (2.5) †34 ab (3.0)35 a (3.5)33 c (1.8)34 ab (4.5)34 (3.4)29–4028–4426–4930–4030–5026–50
PLHT52 b (12.9)59 a (1.9)62 a (18.1)46 c (7.5)58 a (16.8)59 (17.6)25–11025–11823–11325–6020–12020–133
BTN3 b (1.2)3 b (1.3)4 a (1.5)4 a (1.2)3 b (1.3)4 (1.4)1–71–81–92–71–91–9
FLBL212 ab (48.9)214 ab (59.1)229 a (54.2)205 b (33.0)205 b (47.4)222 (53.5)90–32080–38090–370140–280130–34080–380
FLBW21 b (5.5)19 c (6.1)19 c (5.8)24 a (5.1)16 d (5.2)19 (6.0)9–306–307–3014–306–306–30
FLSL76 b (16.4)81 a (17.6)84 a (16.5)70 c (11.1)82 a (13.6)82 (16.6)50–17040–14030–14055–10050–12030–170
PEDL191 ab (71.8)197 ab (67.6)175 bc (58.8)164 c (49.7)206 a (78.4)181 (63.6)50–37060–40015–38055–27070–40015–400
PANEX116 ab (68.8)117 ab (59.4)91 c (53.6)99 bc (54.6)124 a (78.0)100 (59.8)0–27020–3000–2800–3000–3200–320
INFL145 b (41.9)191 a (51.2)209 a (53.5)108 c (16.6)198 a (50.9)193 (58.2)80–370110–40022–35070–140120–38022–400
INF–PBN16 ab (3.5)15 ab (4.8)17 a (4.1)15 ab (3.6)14 b (3.4)16 (4.2)8–245–265–297–257–235–29
DF: days to 50% flowering; PLHT: plant height (cm); BTN: basal tillers number; FLBL: flag leaf blade length (mm); FLBW: flag leaf blade width (mm); FLSL: flag leaf sheath length (mm); PEDL: peduncle length (mm); PANEX: panicle exsertion (mm); INFL: inflorescence length (mm); INF-PBN: inflorescence primary branches number. # CP: compactum; CT: contractum; ML: miliaceum; OV: ovatum; PA: patentissimum. $ Means of races were tested following the Newman–Keuls test [18,19]. Means followed by the same letters are not significant at p ≤ 0.05 and means followed by different letters are significant at p ≤ 0.05. value within parentheses indicate standard deviation.
Table 5. Most diverse pairs of accessions identified in the entire set using Gower’s distance.
Table 5. Most diverse pairs of accessions identified in the entire set using Gower’s distance.
Top Ten Pairs of AccessionsGower’s Phenotypic Distance
IPm 370 and IPm 28060.591
IPm 381 and IPm 27340.586
IPm 381 and IPm 27470.581
IPm 460 and IPm 27470.579
IPm 370 and IPm 27340.578
IPm 2748 and IPm 3810.578
IPm 381 and IPm 28060.575
IPm 2748 and IPm 3700.575
IPm 381 and IPm 27230.574
IPm 362 and IPm 27340.572
Table 6. Average phenotypic distance among races.
Table 6. Average phenotypic distance among races.
Racecompactumcontractummiliaceumovatum
contractum0.269
miliaceum0.2640.266
ovatum0.2290.2740.275
patentissimum0.2660.2540.2670.274
Table 7. Principal component analysis of proso millet germplasm conserved at the ICRISAT genebank, India.
Table 7. Principal component analysis of proso millet germplasm conserved at the ICRISAT genebank, India.
Principal Component (PC)PC1PC2PC3
Eigenvalue3.722.331.28
Percent variance37.19%23.3%12.83%
Trait
Days to 50% flowering0.06−0.410.50
Plant height (cm)0.460.000.28
Basal tillers number0.15−0.23−0.26
Flag leaf blade length (mm)0.44−0.13−0.18
Flag leaf blade width (mm)0.280.16−0.56
Flag leaf sheath length (mm)0.400.050.04
Peduncle length (mm)0.190.580.18
Panicle exsertion (mm)0.080.600.20
Inflorescence length (mm)0.42−0.150.32
Inflorescence primary branches number0.34−0.13−0.27
Table 8. Correlation coefficient among quantitative traits of the entire set and each race of proso millet germplasm conserved at the ICRISAT genebank, Patancheru, India.
Table 8. Correlation coefficient among quantitative traits of the entire set and each race of proso millet germplasm conserved at the ICRISAT genebank, Patancheru, India.
DF #PLHTBTNFLBLFLBWFLSLPEDLPANEXINFL
PLHT
Entire0.259 **
compactum−0.126
contractum0.008
miliaceum0.296 **
ovatum−0.566 **
patentissimum0.187
BTN
Entire0.092 **0.183 **
compactum0.0650.200 *
contractum0.0060.121
miliaceum0.0780.214 **
ovatum0.094−0.024
patentissimum0.211−0.02
FLBL
Entire0.117 **0.656 **0.338 **
compactum0.1010.479 **0.358 **
contractum−0.0010.634 **0.371 **
miliaceum0.0740.686 **0.341 **
ovatum−0.1040.0780.286 *
patentissimum0.377 **0.693 **0.135
FLBW
Entire−0.265 **0.267 **0.118 **0.524 **
compactum−0.1360.446 **0.289 **0.593 **
contractum−0.1640.515 **0.287 **0.709 **
miliaceum−0.273 **0.295 **0.088 *0.556 **
ovatum−0.383 **0.350 *0.353 *0.539 **
patentissimum−0.1470.514 **−0.347 **0.373 **
FLSL
Entire0.0370.597 **0.101 **0.551 **0.410 **
compactum−0.1710.417 **0.0510.423 **0.227 *
contractum0.0230.728 **0.295 **0.637 **0.530 **
miliaceum−0.0060.574 **0.0790.547 **0.544 **
ovatum−0.1460.2490.2050.453 **0.332 *
patentissimum0.1070.601 **−0.0820.548 **0.444 **
PEDL
Entire−0.354 **0.380 **−0.149 **0.089 **0.264 **0.341 **
compactum−0.561 **0.592 **−0.105−0.0110.279 **0.219 *
contractum−0.349 **0.519 **0.0440.270 **0.314 **0.448 **
miliaceum−0.268 **0.381 **−0.137 **0.122 **0.297 **0.405 **
ovatum−0.524 **0.825 **−0.285 *0.0280.0760.277
patentissimum−0.666 **0.297 *−0.2410.0190.447 **0.214
PANEX
Entire−0.368 **0.225 **−0.193 **−0.071 *0.153 **0.087 *0.937 **
compactum−0.552 **0.500 **−0.138−0.1040.234 *0.0750.983 **
contractum−0.394 **0.346 **−0.0640.0590.1540.1840.941 **
miliaceum−0.285 **0.238 **−0.175 **−0.0430.152 **0.141 **0.951 **
ovatum0.0060.476 **−0.284 *−0.142−0.178−0.1010.672 **
patentissimum−0.644 **0.219−0.241−0.0470.357 **0.0440.920 **
INFL
Entire0.327 **0.800 **0.209 **0.638 **0.105 **0.602 **0.148 **−0.021
compactum0.0860.709 **0.270 **0.568 **0.264 **0.500 **0.1430.025
contractum0.190.746 **0.244 *0.686 **0.429 **0.712 **0.14−0.076
miliaceum0.276 **0.838 **0.237 **0.696 **0.252 **0.582 **0.244 **0.086 *
ovatum0.1920.257−0.2470.003−0.1660.0140.313 *0.309 *
patentissimum0.442 **0.654 **0.1570.710 **0.1890.542 **−0.113−0.192
INF_PBN
Entire0.0220.496 **0.185 **0.554 **0.390 **0.350 **0.003−0.115 **0.445 **
compactum0.235 *0.0020.1330.218 *0.077−0.022−0.399 **−0.431 **0.264 **
contractum−0.0470.588 **0.1920.689 **0.566 **0.456 **0.114−0.0660.544 **
miliaceum−0.0350.534 **0.210 **0.578 **0.440 **0.381 **0.092*−0.0220.486 *
ovatum−0.061−0.008−0.0310.250.342 *0.157−0.08−0.2310.219
patentissimum0.1040.601 **−0.0730.427 **0.471 **0.297 *0.062−0.0030.384 **
# DF: days to 50% flowering; PLHT: plant height (cm); BTN: basal tillers number; FLBL: flag leaf blade length (mm); FLBW: flag leaf blade width (mm); FLSL: flag leaf sheath length (mm); PEDL: peduncle length (mm); PANEX: panicle exsertion (mm); INFL: inflorescence length (mm); INF-PBN: inflorescence primary branches number. * Significant at p ≤ 0.05; ** Significant at p ≤ 0.01.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Agriculture EISSN 2077-0472 Published by MDPI AG, Basel, Switzerland RSS E-Mail Table of Contents Alert
Back to Top