Introduction
Peanut is an important oil crop in many developing countries including China, which has the highest peanut production and consumption in the world. Remarkable progress has been achieved in peanut genetic improvement during the past four decades. This has led to improved productivity in China. However, the majority of released peanut cultivars have been susceptible to diseases including bacterial wilt (Wang et al., Reference Wang, Zhang, Zhang and Men2003; Liao et al., Reference Liao2003, Reference Liao, Lei, Li, Wang, Huang, Ren, Jiang and Yan2010). The primary limitation for breeding programmes was the use of accessions with a narrow genetic background and poor disease resistance (Wan, Reference Wan2003). For example, there have been about 300 peanut cultivars developed in China and more than 70% have a genetic linkage with the landraces Fuhuasheng and Shitouqi (Jiang and Duan, Reference Jiang and Duan1998; Sun, Reference Sun1998). The poor utilization of these germplasm resources has been exacerbated by a lack of genetic information for the collections (Sun, Reference Sun1998).
In order to utilize the peanut collection more efficiently, the core collection and mini core collection concepts were developed to capture the genetic diversity and variation of the entire collection by studying in an even smaller number of accessions. The ‘core collection’ (i.e. 10% of the entire collection) represents a manageable size and cost-effective entry point into the collections (Brown, Reference Brown1989) for identifying candidate traits of interest and evaluating genetic diversity (Jia and Li, Reference Jia and Li2004). Therefore, a core collection comprising of 831 accessions has been developed at Tifton, Georgia, USA (Holbrook et al., Reference Holbrook, Anderson and Pittman1993), 1704 accessions at the ICRISAT (Upadhyaya et al., Reference Upadhyaya2003) and 576 accessions in China (Jiang et al., Reference Jiang, Ren, Liao, Huang, Lei, Chen, Guo, Holbrook and Upadhyaya2008). However, the size of these core collections seemed still to be beyond a manageable level for breeders. Therefore, mini core collections (i.e. 10% of the core collection and 1% of the entire germplasm collection) with 184, 112 and 298 accessions have been developed at the ICRISAT (Upadhyaya et al., Reference Upadhyaya, Bramel, Ortiz and Singh2002), US Department of Agriculture/Agricultural Research Service (USDA/ARS) (Holbrook and Dong, Reference Holbrook and Dong2005) and in China (Jiang et al., Reference Jiang, Ren, Zhang, Huang, Lei, Yan, Liao, Upadhyaya and Holbrook2010), respectively. These core or mini core collections have been evaluated for various traits, such as drought tolerance (Upadhyaya, Reference Upadhyaya2005), early maturity (Upadhyaya et al., Reference Upadhyaya, Reddy, Gowda and Singh2006), resistance to peanut root-knot nematode (Holbrook et al., Reference Holbrook, Stephenson and Johnson2000) and reduced level of pre-harvest aflatoxin contamination (Holbrook et al., Reference Holbrook, Wilson and Matheron1998).
The entire collection at the Oil Crops Institute of Chinese Academy of Agricultural Sciences at Wuhan is the largest in China with 6839 accessions (Jiang and Ren, Reference Jiang and Ren2006). It should also be noted that there are many hirsuta accessions in the Chinese native collection compared with the USA and ICRISAT collections (Duan et al., Reference Duan, Jiang, Liao and Zhou1995). The Chinese core and primary mini core collections were effective for evaluating the morphological traits and genetic diversity. However, the size of the primary mini core (298 accessions) is still too large to cost-effectively evaluate some important traits such as bacterial wilt resistance and aflatoxin contamination. The objectives of this study were to develop a mini-mini core collection for capturing the variations in the core collection and the entire collection, and to evaluate the mini-mini core collection for bacterial wilt resistance accessions in order to demonstrate the usefulness of the newly developed mini-mini core collection.
Materials and methods
Peanut germplasm collection
The Chinese peanut core collection consists of 576 accessions (Jiang et al., Reference Jiang, Ren, Liao, Huang, Lei, Chen, Guo, Holbrook and Upadhyaya2008), of which 574 accessions were used in this study. Two accessions were not used in the core collection because of their limited seeds available for phenotyping in the field. Based on the botanical features of the varieties, the 574 accessions selected for this study were divided into subgroups as follows: 57 fastigiata; 235 vulgaris; 35 hirsuta; 210 hypogaea; 37 intermediate types. The 5-year data from 2004 to 2008 were collected from the core collection planted on the farm of Oil Crops Research Institute, Wuhan. The field plot trials were carried out using randomized complete block designs in three replications. Experiment plots were 3.0 m long. Peanut seeds were planted to 40-cm-spaced, three-row plots with a seeding rate of 5 seeds/m row. Field management such as fertilization, chemical spray, irrigation and weed control was performed as required. The following agronomic traits were recorded in each accession from ten randomly selected healthy plants according to the Descriptors and Data Standard for peanut (Jiang and Duan, Reference Jiang and Duan2006): plant height; length of the first branch; number of primary branches; number of secondary branches; number of tertiary branches; total branching number; number of branches with pods; pod number per plant; leaflet length; leaflet width; distance between apical and basilar leaflets; petiole length in the field. After the pods were dried, the traits that were carefully measured and recorded according to the Descriptors and Data Standard were peanut pod yield per plant from the selected plants, pod length and width from the average of ten mature pods, seed length and width of the average of ten mature seeds, 100-pod weight from 100 mature pods randomly selected from each accession, seed weight of 100-pod, 100-seed weight and shelling percentage.
Statistical analysis
Similarity coefficients for phenotypic traits were calculated using the software package NTSYS-PC V.2.0 based on the 21 traits described above. The unweighted pair group method with arithmetic average dendrogram was constructed using similarity coefficient data.
Mini-mini core collection development
The 574 accessions in the core collection were divided into subgroups by botanical types. Cluster analysis was performed by employing the cluster procedure of NTSYS-PC V.2.0 based on the data from 21 morphological characters in each botanical type. The tree procedure (NTSYS-PC V.2.0) was used to delineate the clusters. Approximately 10% of the accessions were randomly selected from each cluster to form the mini-mini core subgroup. At least one accession was included from each cluster even if it had ten or less accessions.
Evaluation of the mini-mini core collection for resistance to bacterial wilt
The accessions included in the mini core collection were grown in the disease nursery in Hongan, Wuhan. These were evaluated for resistance to bacterial wilt caused Ralstonia solanacearum for 5 years from 2007 to 2011 using a random complete block design and three replications. Each accession was grown in three 2-m rows (one plot). The space available for each plant was about 10 cm × 30 cm. For each two test plots, there was one plot of a susceptible accession ‘Ehua No. 4’ planted as a control for susceptibility check. Total plants of each accession were counted and recorded at the seedling stage and at harvest. At the stages of 5% flowering and 80% flowering, the percentage of plants showing bacterial wilt symptoms was recorded. The percentage of healthy plants to total plants (excluding the wilted plants not caused by R. solanacearum) was calculated and used in this study.
Results
Morphological characters of the core collection
Based on the data collected from the 5 years (2004–2008) of performance, a high degree of variation in the core collection of 574 accessions (Table 1) was evidenced for all of the 21 morphological traits. The average plant height was 70.4 cm and ranged from 28.1 to 131.0 cm. The average length of the first branch was 81.5 cm, ranging from 39.6 to 133.0 cm. The average total branch number was 9.5, ranging from 3.9 to 31.2 cm. The average 100-pod weight was 166.7 g, ranging from 77.8 to 304.6 g. The average 100-seed weight was 63.3 g, ranging from 33.8 to 117.0 g, and the average shelling percentage was 74.2%, ranging from 59.9 to 81.0%. A high degree of variation was observed in the number of primary branches, the number of secondary branches, the number of branches with pods, pod number per plant and pod yield per plant (Table 1).
Table 1 Mean, range and variation of the 21 quantitative characters in the mini-mini core collection (2004–2008) and in the core collection (Jiang et al., Reference Jiang, Ren, Liao, Huang, Lei, Chen, Guo, Holbrook and Upadhyaya2008)a
PH, plant height; LFB, length of the first branch; NPB, number of primary branches; NSB, number of secondary branches; NTB, number of tertiary branches; TBN, total branching number; NBP, number of branches with pods; STDV, standard deviation; CV, coefficient of variation; PNP, pod number per plant; LL, leaflet length; LW, leaflet width; DABL, distance between apical and basilar leaflets; PEL, petiole length; PL, pod length; PW, pod width; SL, seed length; SW, seed width; 100-PW, 100-pod weight; SWP, seed weight of 100-pod; 100-SW, 100-seed weight; SP, shelling percentage; PYPP, pod yield per plant.
a t 0.05= 1.960; t 0.01= 2.558. b was significantly different (P< 0.05).
Similarity coefficients among the accessions were calculated for the 21 morphological traits. The average, minimum and maximum similarity coefficients among the 574 core accessions were 0.28, 0 and 0.5, respectively (Table 2). Regarding the botanical types, the average similarity coefficients of accessions within the fastigiata, vulgaris, hirsuta, hypogaea and intermediate types were 0.3047, 0.2841, 0.2560, 0.2619 and 0.2938, respectively. These results demonstrated that there was a high degree of morphological variation in the core collection, but there was a low similarity coefficient and a high variation in hirsuta accessions observed.
Table 2 Comparison of similarity coefficients for the morphological traits between the mini-mini core collection and the core collection among the botanical types
Establishment of the mini-mini core collection
The number of 574 accessions was too large to draw dendrogram clustering practically. Therefore, the 574 accessions in the core collection were stratified by botanical types. The similarity coefficients of morphological traits, and the dendrograms (see Supplementary Fig. S1, available online only at http://journals.cambridge.org) showed that the 57 var. fastigiata accessions could be classified into 7 clusters (see Supplementary Fig. S1(A), available online only at http://journals.cambridge.org), the 235 var. vulgaris accessions into 25 clusters (Supplementary Fig. S1(E), available online only at http://journals.cambridge.org), the 35 var. hirsuta into 8 clusters (see Supplementary Fig. S1(B), available online only at http://journals.cambridge.org), the 210 var. hypogaea into 22 clusters (see Supplementary Fig. S1(D), available online only at http://journals.cambridge.org) and the 37 intermediates into 5 clusters (see Supplementary Fig. S1(C), available online only at http://journals.cambridge.org). About 10% of the random samples from each cluster were selected to form a mini-mini core collection. At least one accession was included from those clusters that had ten or less accessions. Therefore, the mini-mini core collection was established with 99 accessions consisting of 10 var. fastigiata types, 41 var. vulgaris, 8 var. hirsuta, 34 var. hypogaea and 6 intermediate types.
Comparison of morphological characters between the mini-mini core and core collections (Jiang et al., Reference Jiang, Ren, Liao, Huang, Lei, Chen, Guo, Holbrook and Upadhyaya2008)
The morphological characters of the core (Jiang et al., Reference Jiang, Ren, Liao, Huang, Lei, Chen, Guo, Holbrook and Upadhyaya2008) and mini-mini core collections were compared in the fields from 2004 to 2008 (Table 1). Of the 21 characters collected, the averages of 20 traits were not significantly different between the mini-mini core and core collections except for the number of tertiary branches. Furthermore, the ranges of all the 21 characters available in the core collection were captured in the mini-mini core collection. For example, average plant heights were 73.0 and 70.4 cm with ranges from 28.5 to 131.0 cm and from 28.1 to 131.0 cm in the mini-mini core and core collections, respectively. Standard deviations and coefficients of variation of all the 21 characters were slightly higher in the mini-mini core collection when compared with the core collection (Table 1). For example, plant height and coefficients of variation were 26.95 and 25.70%, respectively, in the mini-mini core and core collections.
Based on the similarity coefficient data of accessions for the 21 morphological traits, the average, minimum and maximum similarity coefficients of the 99 mini-mini core accessions were 0.2439, 0 and 0.5, respectively (Table 2). Among the botanical types, the average similarity coefficients of accessions in the fastigiata, vulgaris, hirsuta, hypogaea and intermediate types were 0.3202, 0.2235, 0.2454, 0.2235 and 0.2913 with ranges from 0.1957 to 0.4348, 0 to 0.5000, 0.1522 to 0.3913, 0.0870 to 0.4348 and 0.2174 to 0.3696, respectively. The average and maximum similarity coefficients of the mini-mini core accession were only slightly lower than those of the core accession numerically (Table 2).
Comparison of the morphological characters between the mini-mini core and primary mini core collections (Jiang et al., Reference Jiang, Ren, Zhang, Huang, Lei, Yan, Liao, Upadhyaya and Holbrook2010)
The morphological characters of the primary mini core (Jiang et al., Reference Jiang, Ren, Zhang, Huang, Lei, Yan, Liao, Upadhyaya and Holbrook2010) and mini-mini core collections were evaluated in the fields from 2007 to 2008. Among the 21 characters collected, no significant differences were observed in the average values of 20 characters between the mini-mini core collection and the primary mini core collection, except for the number of tertiary branches (Table 3). Furthermore, the ranges of these 21 characters available in the primary mini core collection were all captured in the mini-mini core collection. The standard deviations and coefficients of variation of 20 characters in the mini-mini core collection were similar to those in the primary mini core collection (Jiang et al., Reference Jiang, Ren, Zhang, Huang, Lei, Yan, Liao, Upadhyaya and Holbrook2010).
Table 3 Comparison of the morphological characters between the mini-mini and primary mini core collectionsa
PH, plant height; LFB, length of the first branch; NPB, number of primary branches; NSB, number of secondary branches; NTB, number of tertiary branches; TBN, total branching number; NBP, number of branches with pods; STDV, standard deviation; CV, coefficient of variation; PNP, pod number per plant; LL, leaflet length; LW, leaflet width; DABL, distance between apical and basilar leaflets; PEL, petiole length; PL, pod length; PW, pod width; SL, seed length; SW, seed width; 100-PW, 100-pod weight; SWP, seed weight of 100-pod; 100-SW, 100-seed weight; SP, shelling percentage; PYPP, pod yield per plant.
a t 0.05= 1.960, t 0.01= 2.558. b was significantly different (P< 0.05).
Assessment for resistance to bacterial wilt in the mini-mini core collection
From 2007 to 2011, the percentage of healthy plants (bacterial wilt symptom-free plants) for the susceptible control Ehua No. 4 ranged from 0 to 22.34% with an average of 12.15%. The percentage of the 5-year average healthy plants among the mini-mini core accessions ranged from 0 to 90.06%. In terms of the botanical types, the percentage of healthy plants for the var. fastigiata, var. vulgaris, var. hirsuta, var. hypogaea and intermediate accessions averaged 23.92, 36.52, 42.72, 28.04 and 24.72%, respectively (Table 4). Significant differences among the botanical varieties were observed for bacterial wilt resistance. There were two accessions with an average of over 88% healthy plants in 5-year field study (2007–2011). Zhh0267 (var. vulgaris) had an average of 88.7% healthy plants and zhh2359 (var. hirsute) had an average of 90.1% healthy plants. Although only eight var. hirsuta accessions were included in the mini-mini core collection, the accessions in var. hirsuta had the highest percentage of healthy plants among the four botanical types.
Table 4 The percentages of bacterial wilt disease resistance and their t values of different botanical accessions
Discussion
In China, peanut core and primary mini-core collections have been developed, each with 576 or 298 accessions from the entire collection of 6839 accessions of cultivated peanut. In this study, we developed a mini-mini core collection with 99 accessions from the core accessions based on the analysis of 21 morphological traits. This effort has led to the identification of two accessions with resistance to bacterial wilt disease, and, thus, demonstrates the potential value and practical application of the mini-mini core collection in breeding.
For developing a mini core collection, it is important to employ an adequate sampling method to conserve the genetic variation in the entire germplasm collection. In the present study, the 574 peanut accessions in the Chinese core collection (Jiang et al., Reference Jiang, Ren, Liao, Huang, Lei, Chen, Guo, Holbrook and Upadhyaya2008) were divided into subgroups by botanical types and then clustered by the similarity coefficients of morphological traits in each botanical type. For the development of the mini-mini core collection, 10% of the accessions from each cluster were selected randomly to preserve the genetic diversity of the accessions in each botanical type. The 99 accessions of the mini-mini core collection cover the entire morphological variation of the 21 phenotypic characters in the core collection (Jiang et al., Reference Jiang, Ren, Liao, Huang, Lei, Chen, Guo, Holbrook and Upadhyaya2008) and primary mini core collection (Jiang et al., Reference Jiang, Ren, Zhang, Huang, Lei, Yan, Liao, Upadhyaya and Holbrook2010). Moreover, no significant differences in variability were observed between the core and mini-mini core collections. Therefore, the mini-mini core subset represented the core collection adequately. The variances and the coefficients of variation in the selected subset are thus expected to be higher than those in the initial collection (Hu et al., Reference Hu, Zhu and Xu2000). In agreement with these expectations, the standard deviations and coefficients of variation of all the 21 characters were 1.04 times higher on average in the mini-mini core collection than those in the core collection. Therefore, the sample size of the mini-mini core collection appears adequate and comparable to US mini core collection with 112 accessions (Holbrook and Dong, Reference Holbrook and Dong2005) and ICRISAT mini core collection with 184 accessions (Upadhyaya et al., Reference Upadhyaya, Bramel, Ortiz and Singh2002).
Establishing a mini core collection was aimed at defining a collection that would be easy to use in research and breeding (Zhang et al., Reference Zhang, Zhang, Wang, Sun, Qi, Li, Wei, Han, Qiu, Tang and Li2011). In this study, no significant difference between the primary mini core (Jiang et al., Reference Jiang, Ren, Zhang, Huang, Lei, Yan, Liao, Upadhyaya and Holbrook2010) and mini-mini core collections was observed on the average of 20 characters evaluated except for tertiary branch number. At about one-third of the sample size, standard deviations and coefficients of variation of the 20 characters in the mini-mini core collection were similar to those in the primary mini core collection. Therefore, the mini-mini core subset is more practical than the primary mini core collection in evaluating important traits. For example, it was feasible to evaluate the mini-mini core collection for bacterial wilt resistance in replicated field plots for several years. This resulted in the identification of two accessions with high resistance. Practical applications of this mini-mini core method have been demonstrated by being included in the whole-genome resequencing project of International Peanut Genome Consortium (PGC) (www.peanutbioscience.com), which will be sequenced at the UC-Davis Genome Center. Based on the distribution of accessions with bacterial wilt resistance in botanical types, it is possible to identify more accessions with resistance in the entire collection of hirsuta accessions. Additional studies are in process to evaluate the mini-mini core collection for other important traits such as resistance to aflatoxin contamination (Liao et al., Reference Liao, Zhuang, Tang, Zhang, Shan, Jiang and Huang2009), seed oil content and fatty acid composition.
Acknowledgements
The authors express their appreciation to Dr Ina Woodruff, Dr John Woodruff and Robert Thomason for reviewing and editing of this manuscript. The authors are grateful to the research grants from the National Basic Research Program of China (2011CB109300), the Chinese National Natural Science Foundation (30571132) and the Crop Germplasm Program (NB05-070401-32). The authors would also like to express their appreciation to the International Peanut Community and PGC for the joint efforts in sequencing peanut whole genome (www.peanutbioscience.com).
