If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
1 Present address: Instituto de Pesquisas Jardim Botânico do Rio de Janeiro, Rua Pacheco Leão 915, Jardim Botânico, 22460-030 Rio de Janeiro, RJ, Brazil.
1 Present address: Instituto de Pesquisas Jardim Botânico do Rio de Janeiro, Rua Pacheco Leão 915, Jardim Botânico, 22460-030 Rio de Janeiro, RJ, Brazil.
Affiliations
Laboratório de Genética e Biologia Reprodutiva de Plantas (LabGen), Instituto Nacional de Pesquisa da Amazônia (INPA), Av. André Araújo 2936, CEP 69083-000 Manaus, AM, Brazil
1 Present address: Instituto de Pesquisas Jardim Botânico do Rio de Janeiro, Rua Pacheco Leão 915, Jardim Botânico, 22460-030 Rio de Janeiro, RJ, Brazil.
Illegal logging is one of the main causes of ongoing worldwide deforestation and needs to be eradicated. The trade in illegal timber and wood products creates market disadvantages for products from sustainable forestry. Although various measures have been established to counter illegal logging and the subsequent trade, there is a lack of practical mechanisms for identifying the origin of timber and wood products. In this study, six nuclear microsatellites were used to generate DNA fingerprints for a genetic reference database characterising the populations of origin of a large set of mahogany (Swietenia macrophylla King, Meliaceae) samples. For the database, leaves and/or cambium from 1971 mahogany trees sampled in 31 stands from Mexico to Bolivia were genotyped. A total of 145 different alleles were found, showing strong genetic differentiation (δGregorious = 0.52, FST = 0.18, GST(Hedrick) = 0.65) and clear correlation between genetic and spatial distances among stands (r = 0.82, P < 0.05). We used the genetic reference database and Bayesian assignment testing to determine the geographic origins of two sets of mahogany wood samples, based on their multilocus genotypes. In both cases the wood samples were assigned to the correct country of origin. We discuss the overall applicability of this methodology to tropical timber trading.
This study examines a method for identifying the source populations of mahogany (Swietenia macrophylla), one of the world's most valuable and widely traded tropical timbers, to the region of origin. Illegal logging and trade in illegal timber and wood products are the cause of many economic and ecological problems in both producer and consumer countries, resulting in the need to track imported timber back to its source. According to estimates, approx. 50% of timber exports from the Amazon Basin, Central Africa, South-East Asia and the Russian Federation originate from illegal logging [
]. Illegal logging is believed to be one of the chief causes of worldwide deforestation, and trade in illegal timber and wood products creates market disadvantages for products from legal and sustainable forestry. The WWF estimates the global costs of illegal timber at approx. €15 billion per year [http://wwf.panda.org/about_our_earth/about_forests/deforestation/forest_illegal_logging/].
In December 2010, new European Union (EU) timber regulations (No. 995/2010) came into force that will make it illegal, from March 2013, to place illegally harvested timber and timber products on the European market. The new rules prohibit the trade of illegally sourced timber and oblige operators to perform due diligence by seeking guarantees that the timber products they sell have been harvested in a sustainable way and according to the laws of the country of origin. To facilitate this, the EU is negotiating Voluntary Partnership Agreements (VPA) with individual timber-producing countries. VPA countries agree to export to the EU only verified legal timber with a FLEGT (Forest Law Enforcement, Governance and Trade) license. The new EU rules are modelled on similar legislation adopted in the USA in 2008, as an amendment of the Lacey Act. This amended act prohibits in the USA all trade in plants and plant products, including timber and timber products, that are illegally sourced from any US state or foreign country, and requires importers to declare the country of harvest and the species name of all plants contained in their products. In addition to legislation that targets the illegal timber trade, the Convention on International Trade in Endangered Species (CITES) aims to regulate international trade in selected animal and plant species to help ensure their survival. Selected species are listed in one of three appendices, depending on the degree of endangerment. With appropriate controls in place, commercial trade is allowed from natural and plantation sources of species listed in Appendix II, such as mahogany species (Swietenia spp.). Since August 2010 the EU has prohibited all imports of S. macrophylla from Bolivia and requested special permissions for imports from all other countries except Mexico and Guatemala. Because of problems with identification, CITES requires trade in look-alike species to be regulated as well. In combination, these regulations are intended to increase pressure on forest concession holders and timber traders to use credible methods to verify the source of their timber.
To meet these objectives, practical mechanisms are needed for identifying the origin of timber and wood products. Falsification of the country of origin is a well-documented problem in the international tropical timber trade. Typically, falsification occurs at the point of import and usually involves the production of false phytosanitary certificates (which verifies inspection and states that a specified disease was not found or does not occur in the area of production), invoices, or certificates of origin. Existing timber tracking systems use external electronic tags or paper-based documentation to trace back the timber origin, both vulnerable to falsification. Another common problem at smaller spatial scales is the false declaration of timber that has been logged outside a registered concession or within a protected area. Thus tools that can validate the point of harvest are also of interest to legitimate enterprises.
In this study, we test a tracking method based on DNA profiling for mahogany (S. macrophylla). DNA fingerprints are inherent natural characteristics of the wood itself and are therefore not falsifiable. We describe the development of a geo-referenced genetic reference database for S. macrophylla that allows wood samples to be traced back to their geographical regions of origin, following genotyping of the sample. The database uses a DNA profile based on six nuclear microsatellite loci (nSSRs) and draws on samples collected from locations across Latin America. The resolution and performance of the database were assessed by blind testing of two sets of mahogany wood samples using a multilocus assignment procedure. The genetic diversity, differentiation and the spatial genetic pattern of the genetic reference dataset were characterised and the statistical power of the assignment test was estimated.
2. Materials and methods
2.1 Species description
Big leaf mahogany (S. macrophylla King, Meliaceae) is the most valuable commercial hardwood species from the Neotropics. The species has a wide geographic range from Mexico through Central America and across the southern Amazon basin of Bolivia and Brazil [
Flexible mating system in a logged population of Swietenia macrophylla King (Meliaceae): implications for the management of a threatened neotropical tree species.
]. It can reach 40 m in height and is considered fast-growing but plantations are often inhibited by the larvae of Hypsipyla grandella, which destroy apical meristems.
Since the late 1700s, S. macrophylla has been heavily exploited [
] and today is threatened throughout its natural range. In Central America populations have declined by 80% in the last 50 years and some are already extinct [
]. In 2003, because of the high risks to population viability associated with over-exploitation and habitat destruction, S. macrophylla was listed on CITES Appendix II [
From previous studies using nuclear microsatellites, S. macrophylla has been shown to have high levels of genetic diversity and significant genetic differentiation among populations in both the Brazilian Amazon [
Population genetic structure of mahogany (Swietenia macrophylla King, Meliaceae) across the Brazilian Amazon, based on variation at microsatellite loci: implications for conservation.
]. More recently, phylogeographic analysis using chloroplast microsatellite markers found contrasting patterns of genetic structure between Central and South American populations [
Chloroplast DNA microsatellites reveal contrasting phylogeographic structure in mahogany (Swietenia macrophylla King, Meliaceae) from Amazonia and Central America.
Genotype data for this study was gathered from a total of 1971 individuals of S. macrophylla from 31 populations in 9 countries across Central and South America (Table 1, Fig. 1). These data were used to construct a genetic reference database (Table 1, Fig. 1). On average 63 (range of 24–400) individuals per population were sampled.
Table 1Location and sample size of Swietenia macrophylla populations used in the study; sample campaign (see description below).
Sort
Country
Population
Sample campaign
Latitude
Longitude
Sample size
1
Bolivia
Bolpebra
1
−11.11
−69.06
92
2
Bolivia
Jorge Cruz
1
−17.98
−63.47
103
3
Bolivia
La Chonta_0
1
−15.73
−62.85
59
4
Bolivia
La Chonta_1
5
−15.78
−62.92
80
5
Bolivia
La Chonta_2
5
−15.78
−62.92
66
6
Bolivia
San Borja
1
−15.01
−67.10
69
7
Bolivia
Sapecho
1
−15.63
−67.17
52
8
Bolivia
Yotau
1
−16.10
−62.98
51
9
Costa Rica
La Cruz
1
11.07
−85.49
50
10
Costa Rica
Los Chiles
1
10.96
−84.74
51
11
Costa Rica
Pocosol
1
10.53
−85.36
49
12
Costa Rica
Sardinal
1
10.09
−84.83
52
13
Guatemala
Peten
2
17.00
−89.66
56
14
Honduras
Colon
2
15.83
−86.00
57
15
Mexico
Veracruz
2
17.66
−94.91
98
16
Mexico
Chiapas
2
16.91
−91.58
52
17
Mexico
Campeche
2
18.41
−89.41
59
18
Mexico
Quintana Roo
2
19.50
−88.08
50
19
Nicaragua
Central
2
13.66
−84.16
53
20
Nicaragua
Terra
2
11.83
−83.91
53
21
Panama
Herrera
2
7.58
−80.66
26
22
Panama
Darien
2
8.83
−78.00
37
23
Belize
East Botes
3
17.64
−88.72
42
24
Belize
Punta Gorda
3
17.62
−88.68
49
25
Brazil
Agua Azul, PA
4
−6.9
−50.27
31
26
Brazil
Cach A, RO
4
−12.5
−61.5
32
27
Brazil
Pontes e Lacerda, MT
4
−15.08
−59.15
24
28
Brazil
Chico Mendes, AC
4
−10.42
−69.3
34
29
Brazil
Cach E, RO
4
−12.57
−61.5
24
30
Brazil
Pimenta Bueno, RO
4
−12.37
−61.43
24
31
Brazil
Marajoara, PA
4
−7.83
−50.27
400
Sample campaigns:
(1) Samples provided from germplasm collection completed in 2000 and 2009 in collaboration with CIAT Bolivia and from a germplasm bank established in the forestry concession of Agroindustrial Forestal La Chonta.
(2) A total of 541 samples from 10 populations were collected from genetic trials in Puerto Rico. These trials had been established in 1965 and 1966 from seeds collected from populations ranging from Mexico to Panama.
(3) Samples from Belize collected during EU-INCO project contract ICA4-CT-2001-10101.
(4) Data was previously obtained from population genetic studies carried out in Brazil
Population genetic structure of mahogany (Swietenia macrophylla King, Meliaceae) across the Brazilian Amazon, based on variation at microsatellite loci: implications for conservation.
Flexible mating system in a logged population of Swietenia macrophylla King (Meliaceae): implications for the management of a threatened neotropical tree species.
Fig. 1Map showing the spatial position of the populations of Swietenia macrophylla. Size of the population marker reflects the sample size for each population.
Two wood collections declared as S. macrophylla were used to validate the DNA profiling method on sawnwood. One was obtained from a timber trader in Hamburg and was comprised of end pieces from each of 20 wooden planks. The second was sent from Bolivia and contained 12 wood pieces from five small joineries in Ascención de Guarayos, Santa Cruz State, Bolivia, about 30 km from La Chonta concession.
4. Genetic analysis
4.1 DNA extraction
The genetic analyses were carried out at two laboratories: the Brazilian samples were analysed at the Plant Genetics Laboratory at INPA, Manaus, AM, Brazil and all other samples at vTI, Grosshansdorf, Germany. At INPA, total genomic DNA was extracted from leaves using a standard CTAB procedure [
Multiplexed systems of microsatellite markers for genetic analysis in mahogany, Swietenia macrophylla King (Meliaceae), a threatened Netropical timber species.
At INPA, microsatellite amplification was carried out in a final volume of 10 μL for single or 25 μL for multiplexed reactions, each containing 1.25–2.0 μM of each forward and reverse primer, 1 unit Taq DNA polymerase, 200 μM of each dNTP, 1× reaction buffer (10 mM Tris–HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl2), BSA (2.5 mg/mL), 5.0 ng of template DNA, and ultrapure water. PCRs were performed in a MJ Research Incorporated PTC 200 thermal cycler under the following conditions for all loci: 96 °C for 2 min, followed by 30 cycles of 94 °C for 1 min, 56 °C for 1 min, and 72 °C for 1 min, and a final elongation step at 72 °C for 30 min. PCR products were added to GeneScan 500 ROX internal size standard (ABI, Inc.), electrophoresed on a 5% denaturing polyacrylamide gel in an ABI Prism 377 sequencer (ABI Inc.) and analysed with Genescan and Genotyper software (ABI, Inc.).
At vTI, PCR was carried out in a final volume of 15 μL for both single and multiplexed reactions, containing 0.1–0.5 μM of each forward and reverse primer, 0.6 unit of Taq DNA polymerase, 200 mM of each dNTP, 1× reaction buffer (80 mM Tris–HCl, 20 mM (NH4)2SO4, 0.02% Tween-20), 1.75 mM MgCl2, 20 ng of template DNA, and ultrapure water. Amplifications were performed using a thermal cycler from Sensoquest using the following program: 94 °C for 3 min, followed by 30 cycles of 94 °C for 30 s, 54–60 °C (depending on annealing temperature of the primers) for 30 s, and 68 °C for 2 min, and a final elongation step at 68 °C for 15 min. PCR products were combined with 8 μL diluted size standard ET-400 ROX (GE Healthcare). Separation and detection of the fragments were carried out on a MegaBACE 1000 96 capillary automated sequencer (GE Healthcare) using Fragment Profiler software version 1.2 (GE Healthcare).
Genotyping of the wood samples was carried out in single PCR reactions with a final volume of 15 μL containing 0.1–0.5 μM of each forward and reverse primer, 0.375 unit Amplitaq Gold 360 DNA polymerase, 200 mM of each dNTP, 1× reaction buffer, 1.75 mM MgCl2), 30 ng of template DNA, and ultrapure water. Amplifications were also performed using a thermal cycler from Sensoquest using the following program for all loci: initial denaturation at 95 °C for 10 min, followed by 45 cycles of 95 °C for 30 s, 54–60 °C for 30 s, 68 °C for 2 min and a final elongation step at 68 °C for 15 min.
4.3 Calibration
A common set of four individuals of S. macrophylla were genotyped for the six nuclear microsatellite loci in both INPA and in vTI laboratories and the results were used to calibrate the genotypes obtained in the two labs. In most cases the fragment sizes were exactly the same within 1 bp. For two loci there was a difference of one base pair and the data were adjusted by a systematic shift of one base pair for all data obtained in the vTI lab.
5. Data analysis
5.1 Genetic diversity and genetic differentiation
Using the program GDA_NT for each locus, the number of alleles (NA), allele frequencies (pij), the observed heterozygosity (Ho), expected heterozygosity (He), and the fixation index (F = 1 − (Ho/He)) were calculated as described by Weir [
] were computed as measures of fixation and genetic differentiation among populations. Genetic distance (D0) was computed to measure the genetic differentiation between pairs of populations [
]. For all pairwise genetic distance (D0) estimations, the genotypes in the two populations were pooled 10,000 times and from this joint group, the two individual populations were reformed by random sampling without replacement. The D0 values were recalculated for both re-sampled populations (i and j). For delta, FST, and GST(Hedrick), the genotypes of the all populations were pooled 10,000 times and from this joint group the 31 (population-level) or 9 (country-level) individual populations were reformed by random sampling without replacement. The differentiation values were recalculated for all re-sampled populations. For each permutation, the genotypic structure of each population represented a random mixture of genotypes from all original populations. The probability of significance was estimated from the proportion of permutations with D0, delta, FST and FST(Hedrick) that were smaller than the observed value. We calculated Spearman's rank coefficient (a non-parametric measure of statistical dependence between two variables) between the geographic and spatial distances among the populations.
5.2 Genotype assignment
Individual assignment tests and group assignment tests were performed using the Bayesian multilocus-approach [
] and GDA_NT (Degen, unpublished). In addition to the parameters calculated with Geneclass 2, GDA_NT computes the probability of exclusion for group assignments. All individuals from the reference data were self-classified to the sampled populations using the leave-one-out approach (self-assignment) [
]. The group was compared with each reference population and assigned to the most probable. There was a most likely population in any reference set, to which the group could always be assigned. However, the set of reference populations might not include the true population of origin for the control group, resulting in a false positive assignment. Therefore, a measure of confidence was needed that the tested individual or tested group truly belonged to a given population [
]. This was achieved by comparing the likelihood value of the test group with the likelihood distribution of groups based on allele frequencies of the population and that were generated from 10,000 sampling runs with replacement. If the observed likelihood of the test group was well outside the 99.9% confidence interval of the distribution then this indicated that the group did not belong to the population.
] for paternity analysis, we used the difference in LOD scores (delta LOD) between the most-likely reference population and the next most-likely reference population as a test criterion for the exclusion probability [
]. In cases where the true population of origin was among the reference samples this difference was clearly larger than when the true population was not part of the reference samples. Again we used the distribution of delta LOD values generated from 10,000 re-sampling runs to locate the observed values.
6. Results
6.1 Genetic diversity
The mean number of alleles per nSSR locus over all populations was 9.13. We observed a total of 145 different alleles. Over all loci, mean expected heterozygosity (He) ranged from 0.61 to 0.83 among the 31 populations (Table 2). There was no significant correlation between the He values and the sample size (Spearman's rank correlation = 0.068, t-test value for hypothesis r = 0 is 0.421, probability > t = 0.6762), indicating that this is a robust measure of genetic diversity insensitive to sample size differences. The Brazilian populations had relatively high values of He while the populations from Nicaragua and Honduras had smaller values. As indicated by significant positive fixation indices, the number of homozygotes in 11 out of 31 populations exceeded Hardy–Weinberg Equilibrium (HWE) expectations, two populations (San Borja in Bolivia, and Colon in Honduras) had a significant excess of heterozygotes, and 17 populations had no significant departure from HWE (Table 2). Aggregating populations at the country level resulted in significant homozygote excess in six out of 9 cases, with one significant negative F-value and only two non-significant values.
Table 2Average expected heterozygosity (He) average fixation index (F), probability of departure from Hardy–Weinberg equilibrium (Prob) for each population used in the study and for populations pooled by country.
Sort
Country
Population
Sample size
He
F
Prob
1
Bolivia
Bolpebra
92
0.710
0.019
0.758
2
Bolivia
Jorge Cruz
103
0.670
0.014
0.632
3
Bolivia
La Chonta
59
0.736
0.032
0.806
4
Bolivia
La Chonta_1
80
0.735
0.003
0.478
5
Bolivia
La Chonta_2
66
0.720
0.086
0.999
6
Bolivia
San Borja
69
0.690
−0.040
0.965
7
Bolivia
Sapecho
52
0.672
0.036
0.792
8
Bolivia
Yotau
51
0.655
0.095
0.990
9
Costa Rica
La Cruz
50
0.610
−0.041
0.887
10
Costa Rica
Los Chiles
51
0.622
0.083
0.982
11
Costa Rica
Pocosol
49
0.823
0.168
1.000
12
Costa Rica
Sardinal
52
0.642
0.155
1.000
13
Guatemala
Peten
56
0.704
0.136
1.000
14
Honduras
Colon
57
0.668
−0.098
1.000
15
Mexiko
Veracruz
98
0.700
0.186
1.000
16
Mexiko
Chiapas
52
0.618
0.192
0.999
17
Mexiko
Campeche
59
0.669
0.017
0.637
18
Mexiko
Quintana Roo
50
0.714
−0.009
0.673
19
Nicaragua
Central
53
0.612
−0.055
0.915
20
Nicaragua
North
53
0.627
−0.041
0.895
21
Panama
Herrera
26
0.611
0.206
1.000
22
Panama
Darien
37
0.722
−0.044
0.904
23
Belize
East Botes
42
0.725
0.019
0.572
24
Belize
Punta Gorda
49
0.706
−0.044
0.912
25
Brazil
Agua Azul, PA
31
0.806
0.015
0.568
26
Brazil
Cach A, RO
32
0.753
−0.020
0.808
27
Brazil
Pontes e Lacerda, MT
24
0.793
−0.003
0.829
28
Brazil
Chico Mendes, AC
34
0.706
0.040
0.817
29
Brazil
Cach E, RO
24
0.795
0.031
0.714
30
Brazil
Pimenta Bueno, RO
24
0.731
0.135
0.995
31
Brazil
Marajoara, PA
400
0.837
0.073
1.000
Belize
All
84
0.718
−0.010
0.687
Bolivia
All
566
0.744
0.088
1.000
Brazil
All
565
0.852
0.101
1.000
Costa Rica
All
202
0.753
0.182
1.000
Guatemala
All
56
0.704
0.136
1.000
Honduras
All
57
0.668
−0.098
1.000
Mexico
All
258
0.743
0.187
1.000
Nicaragua
All
106
0.628
−0.035
0.937
Panama
All
63
0.746
0.136
1.000
Bold numbers indicate significant values with a probaility > 0.95
6.2 Genetic differentiation and spatial genetic structure
All three measures of genetic differentiation and fixation – delta, FST and GST(Hedrick) – had very high and significant values for both differentiation among populations and among countries for the nSSR dataset (Table 3). The differentiation among populations within countries was much smaller and only the GST(Hedrick) for Nicaragua was statistically significant (Table 3).
Table 3Locus-by-locus and mean total genetic differentiation among Swietenia macrophylla populations and among groups of samples pooled by country for nSSRs.
The plot of pairwise genetic distances against spatial distance (km) among populations showed a clear positive correlation (Fig. 2). The high Spearman's rank correlation of 0.824 (t-test value for hypothesis r = 0 is 31.248; probability > t = 0.0000) indicated that genetic distance among populations increased significantly with spatial distance.
Fig. 2Correlation between pairwise genetic distance
], we used all six nSSR loci (sm01, sm22, sm40, sm31, sm32, sm51) to perform self-assignment tests for individuals and groups of individuals randomly selected from the reference dataset. When all 31 populations were used as individual reference populations, 70.7% of all individuals were correctly assigned. The expectation for correctly assigning individuals by chance was only 3.2%. When nine reference populations were created by pooling all populations within each of the nine countries, 82.2% of all individuals were assigned correctly. In this case the expectation of correct assignment by chance was 11.1%. The higher proportion of correct assignments with the populations pooled by country indicated that the blind test of the wood samples would be improved slightly using this strategy. The proportion of correctly assigned samples increased greatly when groups rather than single individuals were assigned. Nine groups of two, three, and four randomly selected individuals were correctly self-assigned at 85%, 90% and 95.6%, respectively, to the reference populations pooled by country.
6.4 Blind test with wood samples
All subsequent analyses were therefore conducted by assigning groups of individual wood samples from the same provider to one of the nine country-level reference populations. Amplification success for the nuclear microsatellites of the wood samples was lower than for the fresh material, resulting in a fragmented dataset, with 25% of the alleles missing for the wood samples from Hamburg and 35% of the alleles missing for the wood samples from Bolivia.
Alleles for at least one nSSR locus were amplified in 15 out of the 20 samples from the Hamburg trader. In three cases identical genotypes were found. Thus we derived 12 distinct multilocus genotypes for the assignment test. These samples were assigned with a high level of confidence (score = 100%) to the Guatemala reference population (Table 4). Exclusion probabilities based on LOD-values were 100% for all nine countries, but the exclusion probability based on delta LOD-values was not significant for Guatemala. Guatemala was confirmed by the timber trader as the tagged country of origin for the wood samples.
Table 4Results of the assignment test for two groups of wood samples of Swietenia macrophylla. Prob 1 = probability of exclusion based on LOD-values, Prob 2 = probability of exclusion based on DELTA LOD-values.
Alleles for at least one nSSR locus were amplified for 9 out of the 11 wood pieces sent from Bolivia. All genotypes were different; so nine different multilocus genotypes were used for the assignment test. These samples were assigned with a high level of confidence (score = 100%) to the Bolivia reference population (Table 4). Exclusion probabilities based on LOD-values were 100% for all countries except for Bolivia which had an exclusion probability of 3% (Prob 1, Table 4). The exclusion probability based on delta LOD-values was also not significant for Bolivia.
7. Discussion
In this study a large, geographically widespread genetic reference database for S. macrophylla was compiled and used to test assignment procedures as well as source blind samples provided by the timber industry. Interest in DNA-based timber tracking has been growing in recent years, as demand and technology make the approach more feasible. Different approaches are being tested. Lowe et al. [
] have reported successful identification of individual merbau logs (Intsia spp.) from both the concession and the saw mill, using a ‘DNA passport’ approach to verify that illegal timber is not being added to shipments between harvest and processing. The alternative used in our study – assignment testing based on reference datasets – requires much more initial investment but is being pursued for several species. To our knowledge, reference data are being compiled and tested for Aquilaria crassna [
7.1 Genetic differentiation and spatial genetic pattern
Strong genetic differentiation was found among the 31 mahogany populations at the nSSRs (δGregorious = 0.52, FST = 0.18, GST(Hedrick) = 0.65) and genetic and spatial distances among stands were correlated. This clear pattern provided a robust basis for successful assignment of genotypes. In general, the accuracy of genotype assignment approaches increases with increasing genetic differentiation among populations [
Population genetic structure of mahogany (Swietenia macrophylla King, Meliaceae) across the Brazilian Amazon, based on variation at microsatellite loci: implications for conservation.
] probably due to the greater geographical scale covered.
7.2 Genotyping wood samples
Genotyping of timber samples required up to six repeats of PCR reactions because the amplification success of some nSSR loci was relatively low (<70%) and allelic drop-out occurred frequent. The amplification success and the detectable fragment size varied from locus to locus, which was probably due to degradation and fragmentation of DNA extracted from timber, as well as the presence of PCR inhibiting substances in the wood [
] also encountered decreased success in amplifying long DNA fragments (100–350 bp) for merbau (Intsia sp.) wood, a timber from the Indonesian region.
From the results of the self-assignment tests, sufficient statistical power (>95% probabilities) was achieved when groups of at least 4 individuals were tested, rather than single individuals. For the blind tests, groups of 9 and 12 individuals were used, which made the assignment robust even when significant allele drop out was present. In most cases, sawn timber will be exported or imported in containers; each container will be filled with planks derived from multiple individuals. The associated trade documents will represent in most cases a claim for the true origin of the whole container load. Thus in practice the necessity for single individual assignment will be more an exception than the rule. It should also be noted that in practice a test will not be searching for the true origin of a timber shipment; instead it will be aiming to verify a match (or mismatch) between the stated origin and the source population. In other words, law enforcement agencies will be aiming to check whether a declared origin is correct or not. This approach is statistically less demanding. As long as the declared origin gets a non significant exclusion probability in the assignment test and has the highest assignment score compared to tested alternatives an authority could accept the statement of origin.
] performed well in the blind tests for S. macrophylla timber origin. Calculation of exclusion probabilities is critical for avoidance of “false positive” assignments. Based on simulated genotypes, we calculated the distribution of LOD-values and compared it with observed values. This yielded a significant exclusion probability for the Guatemalan wood samples, and additional support for the origin of the Bolivian wood samples. It has been reported that this way to compute exclusion probabilities is too conservative [
]. Therefore we also used an approach based on differences between LOD values for the most likely and the second most likely reference population as a test statistic [
]. This less conservative approach gave non-significant exclusion probabilities in both cases. The difference in exclusion probabilities was also affected by the contrasting sample sizes for the reference databases from Bolivia and Guatemala. In Bolivia a total of 566 individuals from eight populations were used for the reference population but only 56 individuals from one population were available for Guatemala.
Acknowledgements
The work was funded by Johann Heinrich von Thünen Institute (vTI) through budget given by the German Federal Ministry of Agriculture, Food and Consumer Protection (BMELV), the UK's Natural Environment Research Council and the EC (via contract ICA4-CT-2001-10101), and the Brazilian Council for Research (CNPq/CT-Amazonia Grant No. 554017/2006-7 to MRL). Access to RBCMA was granted by Programme for Belize. We are thankful for Vivian Kuhlenkamp and Caroline Hartleib for their excellent lab work, Nelson Repollet and Alberto Rodriguez for assistance in collecting the samples in Puerto Rico. We are grateful to Jimmy Grogan and José Ribeiro for field assistance in Brazil. We also acknowledge logistical support provided by the timber companies Serraria Marajoara, Madeireira Juary and Cacique Madeireira in Brazil. And we thank two anonymous reviewers for their helpful comments to improve the manuscript.
References
Pennington T.D.
A monograph of the Neotropical Meliaceae.
in: Pennington T.D. Styles B.T. Taylor D.A.H. Flora Neotropica Monograph 28: Meliaceae. The New York Botanical Gardens,
New York, USA1981
Flexible mating system in a logged population of Swietenia macrophylla King (Meliaceae): implications for the management of a threatened neotropical tree species.
Population genetic structure of mahogany (Swietenia macrophylla King, Meliaceae) across the Brazilian Amazon, based on variation at microsatellite loci: implications for conservation.
Chloroplast DNA microsatellites reveal contrasting phylogeographic structure in mahogany (Swietenia macrophylla King, Meliaceae) from Amazonia and Central America.
Multiplexed systems of microsatellite markers for genetic analysis in mahogany, Swietenia macrophylla King (Meliaceae), a threatened Netropical timber species.
00136-6&id=gr1.jpg){kind=link}
00136-6&id=gr2.jpg){kind=link}