Number of Samples of Each Group of Varieties Originating from Collection Sub-Areas and Rubber Tree Genebank Conservation Centers

Figure 2.2 Rubber leaf development stages and leaf sampling locations for DNA extraction

2.4.2 PCR reaction with SSRs markers for rubber varieties

PCR reactions were performed with three M13-labeled oligonucleotide primers consisting of a strand-specific forward primer with an M13 tail at the 5' end, a strand-specific reverse primer, and a fluorescently labeled M13 primer (Schuelke, 2000). The total volume for the PCR reaction was 10 µL including 5 µL of template DNA (concentration 25 ng/μL), 0.2 µM of each primer pair, 200 µM deoxynucleotides (dNTPs), 2 mM MgCl 2 , 1x PCR buffer, and 1U Taq DNA polymerase. The PCR thermal cycling protocol consisted of the following steps: initial denaturation at 94 o C for 5 min, followed by 10 cycles (Touchdown) with the temperature decreasing by 0.5°C while annealing each cycle from 55°C to 50°C (94°C for 45 s, 55°C for 1 min, 72°C for 1 min 15 s); 25 cycles with annealing at 50°C (94°C for 45 s, 50°C for 1 min, 72°C for 1 min) and a final extension step at 72°C for 30 min. The PCR products were denatured with formamide and subjected to capillary electrophoresis using an ABI 3500 sequencer (Applied Biosystems, Foster City, CA, USA) (Figure 2.3); Data were extracted directly on Excel files with the number and size of polymorphic bands (nucleotide numbers) after analysis using GeneMapper ® v4.1 software. All rubber varieties in the study were analyzed at the CIRAD laboratory (Montpellier - France); the PCR reaction procedure with SSRs markers for electrophoresis using the ABI 3500 sequence reader is presented in Appendix 4.

Figure 2.3 PCR products of 4 samples from the A2387 marker were capillary-generated using an ABI 3500 sequencer and analyzed using GeneMapper software.

2.4.3 Data analysis method

2.4.3.1 Analysis of population genetic parameters

All genetic parameters were analyzed using GENALEX v6.5 software (Peakall and Smouse, 2012); the confidence intervals of the genetic parameters were based on 1,000 bootstraps and the probabilities (P) corresponded to 999 permutations . The formula for calculating the genetic parameters was as follows:

- Observed heterozygosity mean (Ho) = (number of heterozygotes identified)/number of samples;

𝒊

- Average expected heterozygosity or population genetic diversity index (He) He = 1 – ∑ 𝑷 𝟐

Where, Pi is the frequency of the ith allele of the breed group.

- Fixed index of breed groups (F) = (He - Ho) / He = 1 - (Ho/He)

- Inbreeding coefficient (Fst) or genetic differentiation index within breed groups Fst = (Ht - Hs)/Ht

In which, Ht is the expected heterozygosity (He)

Hs is the expected heterozygosity mean (He).

- Polymorphic information index or expected heterozygosity

𝒊=𝟏

𝒊

PIC = 1 – ∑ 𝒏 𝑷𝒂 𝟐

Where, Pa i is the average allele frequency of the ith breed group.

- Principal component analysis (PCA) to detect genetic relationships among rubber variety groups; analysis of molecular variance (AMOVA) to determine genetic variation occurring within the variety sample, between samples and between rubber variety groups.

2.4.3.2 Determining genetic relationships between varieties based on SSRs markers

To detect the genetic relationship between the accessions through the phylogenetic tree constructed using DARWIN v6.05 software (Perrier and Jacquemoud-Collet, 2006). The analysis method is based on the genetic distance from the dissimilarity matrix of the accession data set with a ploidy level of 2 alleles (2 ploidy) with a single pairwise difference index. The analysis of genetic differentiation between accessions was performed with 10,000 iterations; factorial analysis of genetic differentiation according to 5 coordinates and the coordinates were defaulted after factorial analysis; then the phylogenetic tree was constructed according to the method

Neighbor-Joining method from the dissimilarity matrix was analyzed. Genetic differences between the accessions were calculated using the following formula:

d ij = 1 – 𝟏 ∑ 𝑳

𝒎 𝒍

𝑳 𝒍=𝟏 𝝅

Where, dij is the genetic difference between genotypes i and j; L is the number of loci;

π is the ploidy level ( π = 2);

m l is the number of alleles that pair with locus l .

In DARWIN software, the phylogenetic tree is formed containing the value of genetic difference between the samples corresponding to the frequency of appearance that each edge receives on the phylogenetic tree, the value of genetic similarity from 0% to 100% and the samples are grouped together in each genetic cluster.

- A genetic cluster is a collection of all genetically closely related samples that originate from the same branching point on the family tree.

- The variety group is a collection of all rubber varieties collected in the same geographical sub-region of Rondonia state (Brazil) or the same geographical region of the Amazon basin, but not precisely identified in the collection sub-regions.

2.4.3.3 Genetic structure analysis of rubber varieties originating from Rondonia state (Brazil)

The prediction model of the genetic structure of rubber accessions originating from the state of Rondonia was analyzed using two methods.

- Genetic structure based on principal component analysis (PCA)

Genetic structure of accessions from Rondonnia state (Brazil) based on principal component analysis (PCA) using DARWIN v6.05 software. The analysis is based on genetic distances from the dissimilarity matrix with single pairwise index between accessions to produce multidimensional distance graphs of accessions on the coordinate axes.

- Genetic structure by Bayesian clustering method

All 951 rubber accessions originating from the state of Rondonia (Brazil) were subjected to genetic structure analysis using a mixed model with independent allele frequencies among accessions and without providing preferential information on sampling locations, methods of propagation, and accessions.

Bayesian genetic clustering method was performed on STRUCTURE v2.3.3 software (Pritchard et al., 2000), including the following analysis steps:

- The number of genetic clusters (K) that each variety sample was tested for according to the clustering system from 1 to 10 with 20 replicates for each genetic cluster; the burn-in period and the number of replicates were 100,000 and 200,000 respectively for each analysis.

- The special ΔK statistic (Evanno et al., 2005) was used to evaluate the changes in Log probability according to the genetic cluster value (K) and the genetic cluster value

The optimal (K) was determined using the online Structure Harvester program (Earl and vonHoldt, 2012).

- Once the optimal cluster value (K) was determined, using the greedy algorithm from CLUMPP v1.1 software (Jakobsson and Rosenberg, 2007), the sample order was randomly analyzed with 1,000 permutations to estimate the best sample coefficient for each genetic cluster; based on the probability (q) that the sample belongs to the genetic clusters compared with the total number of genetic clusters (K), the sample with probability q ≥ 0.75 belongs to one genetic cluster and the sample with probability q < 0.75 belongs to a mixed genetic cluster. The graph illustrating the number of samples for each genetic cluster was performed using DISTRUCT v1.1 software (Rosenberg, 2004). The subsequent steps of secondary genetic structure analysis were performed using the same method as in the initial Bayesian genetic clustering.

2.4.3.4 Methods for collecting and analyzing latex growth and yield data

- Data collection method

Due to the large number of varieties originating from Rondonia state (Brazil) preserved in Vietnam, the evaluation of agronomic indicators was only carried out on small-scale experiments ( Arboretum, SG ). The experiments were arranged in a completely randomized block design, each experimental plot had 1 tree with 5 replications or 3 trees with 2 replications, all varieties in the experiments were compared with the GT1 control, a clone of the Wickham germplasm. However, in actual production, there are very large differences in the ability to propagate, the rate of grafting and the planning of the area.