American Journal of Botany 99(2): 320–329. 2012.

ULTRA-BARCODING IN CACAO (THEOBROMA SPP.; MALVACEAE)

USING WHOLE CHLOROPLAST GENOMES AND NUCLEAR
RIBOSOMAL DNA1

NOLAN KANE2,5, SAEMUNDUR SVEINSSON2, HANNES DEMPEWOLF2, JI YONG YANG2,

DAPENG ZHANG3, JOHANNES M. M. ENGELS4, AND QUENTIN CRONK2
2
Department of Botany, University of British Columbia, Vancouver BC, Canada V6T 1Z4;
3
SPCL, USDA-ARS 10300 Baltimore Avenue, Bldg. 001 Barc-West, Beltsville, Maryland 20705 USA;
and 4Bioversity International 00057 Maccarese, Rome, Italy

• Premise of study: To reliably identify lineages below the species level such as subspecies or varieties, we propose an
extension to DNA-barcoding using next-generation sequencing to produce whole organellar genomes and substantial nuclear
ribosomal sequence. Because this method uses much longer versions of the traditional DNA-barcoding loci in the plastid and
ribosomal DNA, we call our approach ultra-barcoding (UBC).
• Methods: We used high-throughput next-generation sequencing to scan the genome and generate reliable sequence of high
copy number regions. Using this method, we examined whole plastid genomes as well as nearly 6000 bases of nuclear ribo-
somal DNA sequences for nine genotypes of Theobroma cacao and an individual of the related species T. grandiflorum, as
well as an additional publicly available whole plastid genome of T. cacao.
• Key results: All individuals of T. cacao examined were uniquely distinguished, and evidence of reticulation and gene flow
was observed. Sequence variation was observed in some of the canonical barcoding regions between species, but other
regions of the chloroplast were more variable both within species and between species, as were ribosomal spacers.
Furthermore, no single region provides the level of data available using the complete plastid genome and rDNA.
• Conclusions: Our data demonstrate that UBC is a viable, increasingly cost-effective approach for reliably distinguishing
variet- ies and even individual genotypes of T. cacao. This approach shows great promise for applications where very closely
related or interbreeding taxa must be distinguished.

Key words: cacao; chloroplast; DNA barcoding; evolution; genomics; Gossypium; Malvaceae; plastid; Theobroma; ultra-
barcoding.

Reliable identification of variation below the species level

would provide valuable insight into subspecies ranges, habitat identify subspecies has a long history (e.g., Morin et al., 1992;
Ross et al., 1992; Bardakci and Skibinski, 1994), but is beyond
differentiation, and patterns of gene flow and migration. Addi-
the scope of current widely applied DNA-barcoding efforts by
tionally, it would aid in conservation of important variation
the Consortium for the Barcode of Life (CBOL, 2009).
within species. For example, the conservation status of 2160
subspecies, varieties, or subpopulations of plants and animals To systematically examine variation below the species
has been evaluated by the International Union for level, we propose to sequence and assemble whole organellar
Conservation of Nature (IUCN, 2006), with many high ge- nomes and large (greater than 5 kb) contiguous portions of
profile subspecies listed as endangered, threatened, or extinct. the nuclear genome. We call this ultra-barcoding (UBC; Kane
Reliable identifica- tion of taxa below the species level would and Cronk, 2008) for the simple reason that the approach
also be important for fisheries and wildlife management, and expands the traditional barcoding regions to their full, many-
in breeding efforts to improve crop and lifestock species. The kilobase length. This approach yields a tremendous amount of
use of DNA markers to data for each locus, making it is far more sensitive than
traditional DNA-barcoding (Parks et al., 2009; Nock et al.,
1
2011; Steele and Pires, 2011) and thus may provide the
Manuscript received 6 December 2011; revision accepted 19 December necessary informa- tion to examine variation below the species
2011.
level. The plastid is generally uniparentally inherited (Birky,
This work was funded by a World Bank Development Marketplace
grant awarded to Q.C. and J.E. We thank Chris Grassa, Lambert Motilal, 2001) and behaves as a single nonrecombining locus,
and Loren Rieseberg for helpful discussions and advice; Brian Irish providing a strong signal of population and phylogenetic
(USDA ARS) and Kyle Wallick (USBG) for providing leaf samples; history (Petit and Vendramin, 2007). As such, it is an ideal
Stephen Pinney (USDA ARS) for assistance with DNA extractions; and locus for DNA-barcoding in most taxa. However, evolution is
Jon Armstrong and Jarret Glasscock (Cofactor Genomics, St. Louis) for generally so slow in the plastid genome that there is little
sequencing assistance. Q.C. also acknowledges laboratory support from variation per nucleotide (Palmer, 1985; Wolfe et al., 1987;
NSERC. Zurawski and Clegg, 1987). Thus, the amount of variation
5
Author for correspondence (e-mail: [email protected]) present over short regions may be too low to distinguish
doi:10.3732/ajb.1100570
recently diverged taxa (e.g., Piredda et al., 2011). Because the
plastid is present at many copies per nuclear
American Journal of Botany 99(2): 320–329, 2012; http://www.amjbot.org/ © 2012 Botanical Society of America

320
February 2012] KANE ET AL.—ULTRA-BARCODING IN CACAO 321

genome, plastid genomes can be sequenced relatively inexpen-

sively using low-coverage whole genome shotgun sequence “genetic resource conservation through use”, because local
generated by next-generation sequencing approaches varieties are more likely to be retained by farmers, rather than
(Dempewolf et al., 2010; Meyers and Liston, 2010; Nock et al., be replaced by standard varieties. To realize these benefits,
2011; Straub et al., 2011, 2012). Such a low-pass scan of the however, reliable means of characterizing and identifying ge-
entire genome focuses entirely on the high-copy number netic variation in cacao are important.
fraction of the ge- nome (Steele and Pires, 2011; Straub et al., Variation within cultivated cacao has traditionally been
2012), unlike whole genome sequencing and assembly using divided into three main varietal groups (Cheesman, 1944):
traditional approaches, which typically covers mainly the Criollo, Forastero, and Trinitario. Criollo varieties are consid-
single-copy fraction of the genome. ered to give the best quality product, while Forastero varieties
The UBC approach requires far more sequencing than tradi- are considered to be generally more disease resistant and
tional barcoding, but the cost to generate this additional se- robust (International Cocoa Organization, 2011).
quence has become increasingly affordable with the rise of Hybridization be- tween imported Criollo and Forastero
next-generation sequencing, particularly short-read technolo- varieties in Trinidad and Tobago is the hypothesized origin of
gies such as the Illumina HiSeq. Whole chloroplast genomes the so-called Trinitario varieties that combine the best
and other ultra-barcodes can be inexpensively and accurately characteristics of the parental lineages (Cheesman, 1944).
sequenced using next-generation approaches (Cronn et al., Diverse cacao germplasm is very important for the plant
2008; Parks et al., 2009; Dempewolf et al., 2010; Whittall breeders (Bartley, 2005), and there is a need for markers to
et al., 2010; Doorduin et al., 2011; Steele and Pires, 2011; adequately distinguish accessions (for instance in field
Straub et al., 2011; 2012 Zhang et al., 2011). germplasm banks), for which microsatellites have been
The main focus of this UBC approach in plants is the generally used (Motilal et al., 2009; Irish et al., 2010).
plastid, but nuclear sequence is generated as well and provides Genomic resources for cacao are developing fast. As well
additional useful information. As with the chloroplast, the as complete nuclear genome sequencing (Argout et al., 2010),
ribosomal DNA (rDNA) has the advantage of being there are also two publicly available complete cpDNA genome
multicopy, so requires far less coverage than single-copy sequences, including our Scavina-6 genotype and a genotype
nuclear genes. Additionally, although rDNA is multicopy, it of unknown provenance (Jansen et al., 2011). It is therefore
does not have the same problems of paralogy as other now timely to investigate the feasibility of sequence-based
multicopy nuclear loci such as transposons because the copies approaches in varietal identification, using low-pass whole
do not evolve independently, apparently due to biased gene genome scans.
conversion and unequal crossing over (Arnheim et al., 1980; For these reasons, T. cacao provides an ideal test case for
Coen et al., 1982; Hillis et al., 1991; Wendel and Albert, 1992; the ultra-barcoding concept. While previous researchers have
Elder and Turner, 1995). Termed concerted evo- lution, very successfully sequenced individual whole chloroplast
because the paralogous copies are evolving in concert, this genomes (Meyers and Liston, 2010; Straub et al., 2011) or
property means that rDNA gives strong phylogenetic and even even chloro- plasts and ribosomal DNA (Nock et al., 2011), or
population-level signals (Hillis and Davis, 1988). The used pooled chloroplast DNA samples to assess variation
downside is that if concerted evolution is incomplete or slow, within species (Doorduin et al., 2011), the utility of generating
there may be infraindividual allelic variation at rDNA loci this kind of data for multiple nonpooled individuals within a
(Schaal and Learn, 1988). However, within-individual varia- species has not been tested. In this paper, we use a large data
tion can be dealt with to some extent by generating consensus set of whole ge- nome scans of 10 individuals, in addition to
sequences for each individual, as next-generation sequencing publicly available resources to assess whole-organellar and
approaches are quantitative, providing sequences and copy- whole-ribosomal vari- ation both within T. cacao accessions
numbers for all multicopy rDNA alleles per individual and between T. cacao and a closely related congener.
(Ganley and Kobayashi, 2007).
We recognize that DNA barcoding is conventionally under- MATERIALS AND METHODS
stood to mean delimitation at the species level. However, we
choose to widen the scope to include varietal delimitation, and Sample selection and DNA preparation—Samples for sequencing were
our use of the word hereafter should be understood as such, se- lected to encompass a wide range of cacao variation (Table 1, Appendix
particularly when using the term ultra-barcoding or UBC. As S1; see Supplemental Data with the online version of this article) but focusing
an example of how UBC can be applied, we will here examine on the Trinitario varietal group, which has a complex hybrid origin involving
multiple parental sources of Criollo and Forastero. Leaf material was obtained
whole plastid genomes and ribosomal DNA in Theobroma from ac- cessions held in USDA collections at Beltsville, Maryland, and at the
cacao L. (Malvaceae), a tropical tree crop of great economic Tropical Agricultural Research Station (TARS) in Puerto Rico (Irish et al.,
importance (Wood and Lass, 2001). Cocoa beans are an inter- 2010).
nationally traded commodity used in the manufacture of
choco- late. The enormous range of genetic variation within Illumina sequencing—Total genomic DNA was extracted from fresh leaf
the species (e.g., Motamayor et al., 2008) can be used to tissue using the DNeasy Plant Mini Kit (Qiagen, Valencia, California, USA)
produce chocolate with distinctive flavor profiles. At present, ac- cording to the manufacturer’s protocol. Illumina sequencing libraries were
relatively little choc- olate is marketed with specific varietal pre- pared using standard chemistry and protocols and one lane of paired-end
60- or 80-bp sequence was generated for each sample on Illumina GAII
information, although the use of varieties is increasing due to machines by Cofactor Genomics of St. Louis
connoisseur interest in single variety and known provenance (http://www.cofactorgenomics.com/).
chocolates. However, a rise in interest in cacao varieties is
likely to have positive impacts on the genetic diversity in the De novo assembly of chloroplast and rDNA—De novo assembly of refer-
production systems. First, it is likely to enhance the product ence T. cacao chloroplast and rDNA was accomplished using approaches de-
value chain, especially to the economic benefit of small scribed in Dempewolf et al. (2010), for the Scavina-6 genotype. Briefly, reads
producers. It is also likely to have positive impacts for were trimmed and cleaned, and assembled using the ABySS (Simpson et al.,
conservation under the paradigm of 2009) and SOAPdenovo (http://soap.genomics.org.cn/soapdenovo.html)
assemblers. Chloroplast-encoded contigs were identified using NCBI program
blastn (Ca- macho et al., 2009), with a minimum e-value of 10 −20 and
322 AMERICAN JOURNAL OF BOTANY [Vol. 99
minimum sequence identity of 85% to Gossypium hirsutum chloroplast
genome (Lee et al., 2006),
TABLE 1. Provenances of Theobroma material sequenced.

Name Source of material Notes Accession

EET-64 (T. cacao) USDA (TARS), Puerto Rico Hybrid between Upper Amazon Forastero (Nacional from Ecuador)
and Trinitario (from Venezuela). PI 275669
Criollo-22 USDA (SPCL), Beltsville, MD Pure Criollo variety Criollo-22
Stahel (T. cacao) USDA (TARS), Puerto Rico Trinitario with similarities to lower Amazon Forastero MIA 27956
Pentagonum (T. cacao) USDA (TARS), Puerto Rico Trinitario (Criollo-type) TARS 12044
Scavina-6 (T. cacao) USDA (SPCL), Beltsville, MD Upper Amazon Forastero, Peru MIA 29885
Amelonado (T. cacao) USDA (TARS), Puerto Rico Lower Upper Amazon Forastero TARS 16542
ICS39 (T. cacao) USDA (TARS), Puerto Rico Trinitario TARS 16664
ICS06 (T. cacao) USDA (TARS), Puerto Rico Trinitario TARS 16658
ICS01 (T. cacao) USDA (TARS), Puerto Rico Trinitario TARS 16656
T. grandiflorum (Cupuaçu) USDA (TARS), Puerto Rico Species related to T. cacao. Wild and cultivated in Amazon Basin 04-0254
Notes: MD: Maryland, USA; TARS, Tropical Agriculture Research Station; SPCL, Sustainable Perennial Crops Laboratory; USDA, U. S. Department
of Agriculture

which was the most closely related published full chloroplast genome at the
HQ244500 for the de novo assembly, and GenBank accessions JQ228379–
time of the analysis. The most complete chloroplast assemblies resulted from
JQ228389 for the reference-guided assemblies). The rDNA sequences are
SOAPdenovo, but both assemblies were aligned to the published chloroplast
GenBank accession JQ228369–JQ228378.
genome of G. hirsutum (Lee et al., 2006). Adjacent contigs were merged when
overlap was greater than 15 bp, and shorter overlaps and small gaps were
filled using custom perl scripts with trimmed Illumina read data as in Phylogenetic analysis—An alignment of the rDNA and the plastid se-
Dempewolf et al. (2010). Quality of the final assembly was confirmed and quences (our 10 samples as well as an additional sample of T. cacao, GI
error-corrected by aligning the quality-trimmed Illumina reads using the 328924764), which included all SNPs observed in the data nset was made
program MOSAIK (see below) for each genotype. Similarly, to generate a using the program MUSCLE (Edgar, 2004) and used for phylogenetic analy-
reference rDNA, assemblies were blasted against the existing rDNA ses. A phylogeny for the plastid data was analyzed under maximum likeli-
sequence for T. cacao (GI:7595560, GI:27447189, GI:133854413) and hood (ML; Felsenstein, 1973) using the program Garli 2.0 (Zwickl, 2006).
related species (GI:19919576–19919578; The TIM base substitution model was chosen for the ML analysis, based on
Soltis et al., 2003). Adjacent contigs were again merged when overlap was model testing with the program jModeltest 0.1.1 (Guindon and Gascuel,
greater than 15 bp, and gap filling and extension was performed where 2003; Posada, 2008). One thousand bootstrap replicates were performed to
possible using the unassembled Illumina reads. obtain support values for the phylogeny. The majority rule consensus tree
The full-length chloroplast sequence was annotated using DOGMA (Dual was generated and tree drawn and rooted using the program Figtree v1.3.1
Organellar GenoMe Annotator; Wyman et al., 2004), with additional informa- (http://tree.bio.ed.ac.uk/software/figtree/). The rDNA data set was analyzed
tion about splice sites and open reading frames provided from comparisons with a network-based approach, as rDNA undergoes recombination and is
with another completely sequenced T. cacao plastid genome (Jansen et al., therefore likely to violate the assumption of evolving in a bifurcating manner.
2011) as well as the fully annotated Gossypium chloroplast genomes (Ibrahim The program SplitsTree4 (Huson and Bryant, 2006) was used to visualize the
et al., 2006; Lee et al., 2006). The completed annotation was illustrated using relationships among the T. cacao varieties and to T. grandiflorum and were
OGDraw (Organellar Genome Draw; Lohse et al., 2007). displayed as a phylogenetic network. Support values at each node in the
network were estimated by running 1000 bootstrap replicates.
SNP genotyping for each sample—Trimmed, cleaned paired-end reads
were mapped to the reference T. cacao chloroplast and rDNA sequence using Comparisons with Gossypium—Overall structural similarities between the
MOSAIK (Hillier et al., 2008), with a hash size of 12, 12 mismatches allowed, full-length plastid genomes of Theobroma and Gossypium (Lee et al., 2006;
and an ACT score of 35. Sorted alignment files were converted to BAM Ibrahim et al., 2006) were examined aligning the two genomes with the blastz
format and single nucleotide polymorphisms (SNPs) were called for each algorithm (Schwartz et al., 2003) and illustrated using the program zPicture
sample against the reference using the program SAMTOOLS (Li et al., 2009), (Ovcharenko et al., 2004).
with a minimum quality of 20 for SNPs and 50 for insertions and deletions.
Addition- ally, alignments were manually checked and confirmed with the
raw reads for all significant differences from the reference. RESULTS
Sequence accession numbers—All of our WGS Illumina/Solexa reads are Illumina sequencing—We obtained between 1.7- and 4.6-
available from NCBI’s Short Read Archive (accession SRA048198.1), with
our annotated whole plastid genome assemblies also uploaded (GenBank Gbp sequence per sample after removing low-quality reads,
accession

TABLE 2. Illumina sequence summary statistics and observed average coverage of the nuclear and chloroplast genome for Theobroma based on Burrows-
Wheeler Aligner (BWA) alignments (see text).
Read length No. of pairs of No. of pairs after Nuclear genome Chloroplast genome rDNA
Name (bp) reads filtering Total bp coverage coverage coverage
EET-64 (T. cacao) 60 3.3E+07 3.2E+07 3.9E+09 9.36685 871 2669.9
Criollo-22 (T. cacao) 60 2.3E+07 2.1E+07 2.5E+09 6.08258 850.1 900.1
Stahel (T. cacao) 60 3.0E+07 2.8E+07 3.4E+09 8.17719 1229.1 2377.6
Pentagonum (T. cacao) 80 2.9E+07 2.6E+07 4.1E+09 9.87637 1539.8 4437.4
Scavina-6 (T. cacao) 60 1.9E+07 1.4E+07 1.7E+09 4.18122 186.9 783.9
Amelonado (T. cacao) 60 3.5E+07 3.4E+07 4.1E+09 9.89147 1291.4 3945.4
ICS39 (T. cacao) 80 3.2E+07 2.8E+07 4.5E+09 10.7130 1117.6 2735
ICS06 (T. cacao) 80 3.2E+07 2.8E+07 4.6E+09 10.9555 1324.7 3269.8
ICS01 (T. cacao) 60 2.5E+07 2.4E+07 2.9E+09 7.01850 944.6 3043.6
T. grandiflorum (Cupuaçu) 60 3.6E+07 3.4E+07 4.1E+09 9.86942 772.5 2662.2
or 4.2–11× coverage of the nuclear genome per sample
based on an estimated genome size of 416 Mbp (Figueira et distinguish any of the major varieties of T. cacao, with only
al., 1992). This was more than adequate coverage to assemble rbcL showing any variation and that only at one nucleotide.
both plastid and rDNA: 187–1540× average coverage of the The most variable 500-bp region of the chloroplast within T.
plastid and 784–3946× average coverage of the rDNA (Table cacao was in the 3′ end of the ccsA gene, with four SNPs
2). within T. cacao and two SNPs between T. cacao and T.
grandiflorum. When using the entire plastid genome, however,
De novo assembly of chloroplast and rDNA—Our de novo we observed orders of magnitude more vatiation: 78 SNPs
chloroplast assembly covered the entire 160 546-bp circular segregating within
plastid genome (NC_014676, Fig. 1). Sanger sequence of T. cacao and 415 SNPs observed between species. Indeed,
2162 bp of the plastid genome confirmed the assembly in nine there were multiple high-quality SNPs uniquely distinguishing
re- gions (GI: 338190271–338190279), and also confirmed the each sample sequenced with the exception of EET and Criollo-
length of repeats in microsatellites (Yang et al., 2011). The de 22, which were identical. The rDNA showed similarly high
novo rDNA assembly spanned the entire expressed portion in- varia- tion, particularly within ITS1, ITS2 and the ETS (Table
cluding ETS, 18S, ITS1, 5.8S, ITS2, and 28S (Fig. 2), for a 4). Again, however, the variation when including the entire
total of 5826 bp. rDNA sequence was far greater than any one region, enabling
unique identification of each individual sequenced, this time
SNP genotyping for each sample—Numerous SNPs were including EET and Criollo-22, which differed at several sites.
found both within and between Theobroma species (Figs. 1, 2;
Tables 3, 4), with far more variation between than within spe- Phylogenetic analysis—The phylogenetic analyses revealed
cies for most regions. Three SNPs were confirmed in 95 indi- significant divergence between the major clades of T. cacao.
viduals using Sanger sequence. Each of the examined The ML tree inferred by Garli (Fig. 3) showed two strongly
canonical barcoding regions showed substantial variation supported clades within T. cacao corresponding to the origin
between T. cacao and T. grandiflorum (Table 3), but none had of the Forastero and Criollo varieties. Furthermore, the tree
enough variation to shows that maternal lineages of Trinitario samples come
from both

Fig. 1. Map of the annotated circular chloroplast for Theobroma cacao (outer circle). Inner circle: SNPs segregating within T. cacao. Middle circle:
SNPs fixed between T. cacao and T. grandiflorum.
 Fig. 2. Schematic diagram of the Theobroma cacao ribosomal DNA, including ETS, 18S, ITS1, 5.8S, ITS2, and 28S, with SNPs indicated by
diamonds.

Forastero and Criollo backgrounds. Additional clades are also

well supported within the two main branches of T. cacao chlo- due to difficulties in amplifying matK (Kress and Erickson,
2008; CBOL, 2009) likely due to secondary structure, and this
roplast genomes. The T. cacao sample sequenced by Jansen et al.
has so far only been partially resolved by optimizing PCR
(2011), which has no publicly available provenance information,
prim- ers and protocols (CBOL, 2009).
groups strongly with the Forastero cpDNA types, so is most
likely Forastero or Trinitario. Despite these difficulties, DNA barcoding has been quite
successful for a number of applications. Barcoding has been
The rDNA SplitsTree4 network showed similar patterns to
shown to be a useful tool for identifying species in
the phylogenetic tree, with Forastero and Criollo clustering
biodiversity hot-spots (Lahaye, 2008), identifying material
into distinct clades, and Trinitario intermediate (Fig. 4).
from species listed under the Convention on International
Evidence of substantial gene flow can be observed between
Trade of Endan- gered Species (CITES) appendixes for
the clades as well, particularly in the Trinitario samples.
conservation or regu- latory purposes (Lahaye, 2008; Dexter,
et al., 2010; Yesson, et al., 2011), revealing cryptic species
Comparisons with Gossypium—Over 98% of the Gossyp-
(Herbert et al., 2003; Lahaye, 2008; Ragupathy et al., 2009;
ium and Theobroma genomes aligned well, with 94% of the
Newmaster and Ragupathy, 2010), vouchering specimens
bases identical overall. The only major differences between
(Dugan et al., 2007), delimiting species (Muellner et al.,
the genomes is a large insertion between ycf3 and trnF. The
2009), confirming endangered species identification for
other large regions of nonalignment include the intergenic
reintroduction programs (Rowntree et al., 2010), and
region be- tween ndhC and trnV, the longest at nearly 1 kb,
identifying invasive species (Bleeker et al., 2007).
and numerous smaller, mainly intergenic regions. In contrast,
Ethnobotanical research can be greatly aided by this approach
the inverted re- peat regions are much more similar (99%
to species identification as well (Ragupathy et al., 2009; New-
identical) over almost 24 kb.
master and Ragupathy, 2010). Important applications in foren-
sic botany (Ferri et al., 2009; Ward et al., 2009), medicine
DISCUSSION (Howard, 2010; Lou et al., 2010), and consumer protection
(Kress and Erickson, 2008) have also been proposed. These
Advantages of ultra-barcoding—Taxon identification by and other potential applications make it clear that DNA
means of DNA sequencing (commonly called DNA barcoding is a useful tool.
barcoding) uses hundreds of base pairs of sequence from an Here we assess a related approach using much larger ver-
organellar ge- nome (mitochondria in animals, plastids in sions of the traditional DNA-barcoding loci to examine within-
plants) or nuclear locus ITS (widely used in fungi; Bellemain species variation. For the identification of haplotype groups
et al., 2010). The standard approach in animals involves within a species, such as for the characterization of crop plant
sequencing a 650-bp portion of the mitochondrial diversity (e.g., Coart et al., 2006), a large amount of sequence
cytochrome oxidase I (COI), a region that contains may be advantageous. The total length of the plastid genome
substantial variation but highly conserved primer sequences provides an upper limit for the amount of sequence that can be
(Hebert et al., 2003). This makes DNA barcoding relatively interrogated from this genomic component, and thus the maxi-
straightforward in most animal taxa, but progress was initially mum amount of data that can be obtained for that locus.
stymied in plants because of the difficulty of identifying a The results reported here demonstrate the benefits of ultra-
high-information region with conserved flank- ing sequences barcoding for application below the species level. While tradi-
appropriate for primer design (CBOL, 2009). tional barcoding often struggles to reliably differentiate
Several alternative chloroplast-encoded barcoding regions species, our plastid and rDNA data distinguish every
were proposed for plants (Pennisi, 2007), including portions of individual cacao plant sequenced, each with its own unique
several protein-coding genes (matK, rbcL, rpoB, and rpoC1) SNP pattern. Fur- thermore, there was much higher
and intergenic spacers (atpF-atpH, trnH-psbA, and psbK- divergence between species
psbI). Recently, a two-locus combination of rbcL and matK
was cho- sen by the Consortium for the Barcode of Life based
on a num- ber of factors including ease of amplification and TABLE 3. Variation in single nucleotide polymorphisms (SNPs) within
sequencing and the ability to discriminate between closely Thoebroma cacao (column 2) and fixed SNP differences between T.
cacao and T. grandiflorum (column 3) in four plastid regions.
related species (CBOL, 2009). The reliance on coding regions
rather than in- trons or intergenic spacers helps to avoid
problems with se- quence quality and alignment. However,
these coding regions have lower variation (and thus lower
ability to discriminate spe- cies) than some intergenic regions.
Moreover, difficulties re- main even when using these
established regions, for instance
Region Within T. cacao Between Theobroma species
psbA-trnH 0 7
rbcL 1 4
matK 0 1
ccsA 4 2
Entire plastid genome 78 415
TABLE 4. Variation within Theobroma cacao and fixed differences between
T. cacao and T. grandiflorum in six rDNA regions.
Region Within species Between species
ETS 3 2
18S 1 0
ITS1 5 6
5.8S 1 1
ITS2 6 5
28S 3 14
Entire rDNA 19 27

of Theobroma, with several hundred bases differentiating spe-

cies across the entire plastid genome for any interspecific com-
parison, while there were none or a handful of differences in
the smaller barcoding regions. Similarly, sequencing of the
major- ity of the chloroplast genome in 17 individuals of
Jacobaea vulgaris showed that even with low levels of
chloroplast se- quence variation per base, the entire chloroplast
provides the potential for a very substantial data set even using
pooled sam- ples (Doorduin et al., 2011). That study identified
somewhat fewer segregating SNPs (32 in total) within the
species, al- though concluded that this may be an
underestimate due to low coverage of many regions of the
genome. UBC also reveals a considerable number of plastid
microsatellite loci (described in Yang et al., 2011), which may
be a useful tool to examine large numbers of individuals.
Plastid haplotype data, due to their nonrecombining nature,
are particularly powerful in distinguishing phylogeographic
pat- terns (Soltis et al., 1997; Cavers et al., 2003), such as those
evident in crop wild ancestors and traditional landraces. In our
samples, for example, the fact that Criollo-22 and EET-64
have 100% identical plastid genomes suggests a very recent
common ances- tor, presumably a maternal grandmother or
great-grandmother. Records show that one parent of EET-64 is a
Venezuelan Trini- tario, ‘Venezuelan Amarillo’, which is a
hybrid between Cri- ollo and Lower Amazon Forastero. Our
results characterize the maternal lineage of that
hybridization event and confirm that some of the Ecuadorian
Arriba cacao have a Criollo maternal pedigree, which was
inherited through the Venezu- elan Amarillo.
Broader phylogeographic studies in plants may be limited
by the lack of variation in standard plastid loci (Schaal et al.,
1998). The short sequences used as standard barcoding loci
mean that phylogeography is generally considered to be be-
yond the scope of traditional DNA barcoding. This impedi-
ment is largely removed by ultra-barcoding. Moreover,
where hybridization is rife, such as between crop lineages or
between crops and their wild relatives, a haplotype approach
only tells part of the story, so the ability of ultra-barcoding Fig. 3. Phylogenetic tree representing relationships of sequenced
to provide high-resolution markers for tracking maternal in- Theobroma cacao genotypes and T. grandiflorum, based on complete
heritance as well as nuclear loci subject to biparental inheri- plas- tid genomes for each individual (−lnL = 221 170.271). Sample
tance can be of considerable assistance in plant breeding and names are appended with variety information: (C) Criollo, (F) Forastero,
diversity studies. As sequencing costs fall, it is likely that (T) Trinitario.
Sanger-based sequencing of a few plastid loci will be in-
creasingly replaced by routine whole-plastid sequencing of
multiple individuals. ITS2), have been a reliable workhorse for species
Another advantage of ultra-barcoding is the generation of identification and phylogenetics (Yao et al., 2010). However,
both nuclear as well as plastid data. The low-depth total ge- these spacers are often limited by their short length and low
nomic DNA scans that generate whole plastid data also variation. Ultra- barcoding readily generates much longer
reliably sequence high copy-number regions of the nuclear ribosomal repeat re- gion sequences, including both highly
genome, such as the ribosomal repeat region. Two subregions conserved and highly variable regions, considerably expanding
of the ribosomal repeat, the internal transcribed spacers the number of nu- clear polymorphisms detected in our
(ITS1 and analysis. Unfortunately the
 Fig. 4. SplitsTree4 split decomposition network of the rDNA consensus sequence for each individual sequenced. Only bootstrap values above 50 are
shown. Sample names are appended with variety information: (C) Criollo, (F) Forastero, (T) Trinitario.

most variable region of the ribosomal repeat unit, the

Our data indicate that low-pass genome scans can identify
intergenic spacer (IGS), is presently proving difficult to work
varieties even when a recent ancestor is shared. The plastid
with. This is likely to be due to the rapid evolutionary rate of
SNP variation is sufficient to uniquely identify nearly every
this region and the existence of multiple variants, particularly
individ- ual examined, as well as showing clear distinctions
length polymorphisms (Ambrose and Crease, 2011). The
between the major varieties. The chloroplast genomes showed
variation within individuals is to be expected in rapidly
a clear phylo- genetic signal as would be expected of a
evolving re- gions as there will be insufficient time for
nonrecombining re- gion (Fig. 3). Interestingly, there is good
concerted evolu- tion to homogenize copies after they arise.
support for a separation of the Criollo and Forastero varieties,
However, even excluding this region, there is ample SNP
with the Trinitario variet- ies being scattered through the tree,
polymorphism to examine the comparative evolutionary
again, as would be expected from a mixed group of hybrids
history of nucleus and plastid.
(Fig. 3).
Practical barcoding applications in cacao—Given the re- The data from the rDNA is sufficient to uniquely identify
sults reported here, it is possible to ask how well a whole- each accession. With these data, however, a less strong hierar-
genome scan technique works for practical applications and chical structure is recovered as is to be expected from a recom-
also to assess the prospects for future work. There is consider- bining nuclear region. We therefore chose to represent this by
able need to identify and characterize cacao germplasm for nu- a network graph (Fig. 4). Remarkably, though, there is also a
merous purposes including breeding work, industrial applications sug- gestion of a separation between Criollo and Forastero
and efficient genetic resource conservation. Accurate varietal varieties. In Fig. 4, Criollo and Forastero varieties are widely
identification can help to support the “value chain”, from con- separated on the graph with Trinitario varieties in between,
sumer to grower, with the aim that local growers obtain increased consistent with their hybrid origin.
economic benefits. Similar considerations are also arising from Previously, nuclear microsatellites have been used with suc-
increasing niche marketing of other crops, such as coffee cess for germplasm identification in cacao (Irish et al., 2010;
(Fitter and Kaplinsky, 2001). The use of genetic Motilal et al., 2009). Sequence data offer some advantages as
technologies to support the value chain is important in other an alternative. The lower mutation rate of SNPs means that
bioproduct indus- tries such as the meat industry (Sosnicki and they are less prone to recurrent mutation, back-mutation and
Newman, 2010). conse- quent homoplasy. SNPs are also preferable for
There are also expected conservation benefits from increased measuring popu- lation genetic parameters (Kronholm et al.,
growing of local varieties when supported by accurate varietal 2010; Whitlock 2011). Furthermore, the ability to
identification. This “conservation-through-use” paradigm simultaneously gather infor- mation on variation in two
(Coomes, 2004; Newton, 2008) is an important aspect of genetic genomes (plastid and nuclear) offers great advantages in
resource conservation for plants of economic value, besides the interpreting patterns of ancestry and parent- age for breeders.
value of higher genetic diversity in the production system to in- The ribosomal DNA information gap—Repetitive
crease resilience, e.g., in times of change. sequences in general are given short shrift in full-length
genomes, with
full-length rDNA repeats not present in most publically avail- has nu- merous advantages, including higher resolution than
able “fully sequenced” genomes. For instance, neither of the smaller
two released T. cacao genomes (GenBank CACC01000001–
CACC01025912, Argout et al., 2010, and http://www.cacaog-
enomedb.org/) has even a single full rDNA repeat, but rather
only very small portions on several chromosomes. This is
likely due to difficulty of assembly due to polymorphism
among cop- ies (Straub et al., 2011) as well as a focus on the
single-copy portion of the genome. The question remains as to
whether the IGS is a treasure trove of information waiting to
be tapped or whether the amount of infra-individual variation
is so high (e.g., Chou et al., 1999) that the problems of
assembly, alignment, and analysis are intractable (Rogers and
Bendich, 1987). These problems with the IGS are all
potentially solvable, with better algorithms, better technology,
and better approaches. For the IGS, longer reads such as 454’s
new chemistry may resolve some of these problems. For the
moment, however, the varia- tion present at the IGS is
tantalizingly out of reach for most taxa. Similarly, while whole
plastid genomes are available for 126 angiosperm species and
subspecies as of June 2011, with more available nearly
weekly, only 22 angiosperm mitochon- dria are available,
most being clustered within a handful of families with nearly
half (45%) being from the Poaceae alone. As more genome
projects involve major evolutionary compo- nents, we hope
that the gaps in rDNA and mitochondrial se- quence will be
filled.

Technology and cost—Cost per megabase of data is

dropping rapidly. This is particularly true for the new HiSeq
chemistry, where reagent costs are less than US$0.04/Mb,
although many other technologies are close to this value
(Glenn, 2011). With expected yield of close to 40 Gb/lane
(Glenn, 2011), however, the standard 12 Illumina barcodes
(http://www.illumina.com) are far too few, yielding much
more coverage than necessary for an ultra-barcoding approach
in most species. However, with the use of 48 barcodes, a 0.5-
1Gb genome would have ~1× coverage, more than enough for
UBC (Straub et al., 2012). Our hope is that major work is
given to low-cost chemistry and protocols for library
preparation. A combination of improved and more affordable
robotics and low-cost reliable chemistry for library preparation
are the key to making these techniques cost-effective for the
widest-possible range of applications. Manufacturers are
working in this direction, with 24-plex indexing Illumina
DNA sample prep kits available soon
(http://www.illumina.com/sup- port/faqs.ilmn), and 48-level
indexing validated primers already available from other
vendors (e.g., http://www.biooscientific. com). With current
commercial costs of 48-plex indexing and library preparation
reagents below US$42/sample (e.g., http://
www.biooscientific.com), and roughly US$1500 in
sequencing costs, a lane of 48 individuals could be run for
~US$3500 (US$73/ sample) in reagent costs. While still more
expensive than a single Sanger sequence per individual
(typically only a few dollars per sequence, or less for some at-
cost facilities), the vast increase in data quantity and ability to
distinguish lineages within species may make the extra
expense of UBC worthwhile for many ap- plications even at
today’s prices. As next-generation sequencing prices continue
to decline, it is clear that the UBC approach will become
increasingly attractive.

The future of DNA barcoding— UBC is a promising

approach, resolving many of the difficulties with current
widely applied DNA-barcoding approaches. This approach
regions, and universality, as the need for taxon-specific primers
is avoided. Nuclear loci such as ribosomal DNA provide sub-
stantial additional information that makes this approach even
more powerful. Additional loci such as the mitochondrial ge-
nome, nuclear microsatellites, LTRs, TEs, and other repeats
could also be assayed with this approach. For numerous breed-
ing, medicinal, consumer protection, and forensic botany ap-
plications, it would be highly useful to identify samples to below
the species level. Additionally, numerous questions related to
phylogeography, population genetics, phylogenetics, and mo-
lecular evolution would benefit greatly by this type of data set.
Finally, we should emphasize again that UBC does not remove
the need for continued use of traditional barcoding methods
currently in use, but rather provides data below the species
level. The main stumbling blocks for UBC are cost and the
somewhat higher quality and quantity of DNA required, as
well as the bioinformatic and computational resources needed
to deal with large amounts of next-generation sequence data.
Nev- ertheless, as technology and methodologies continue to
improve rapidly, we believe that it will soon be practical to
sequence and assemble whole plastid genomes for a broad
range of applications.

LITERATURE CITED
AMBROSE, C. D., AND T. J. CREASE. 2011. Evolution of the nuclear ribo- somal
DNA intergenic spacer in four species of the Daphnia pulex complex.
BMC Genetics 12: 13.
ARGOUT, X., J. SALSE, J. M. AURY, M. J. GUILTINAN, G. DROC, J. GOUZY,
ET AL. 2010. The genome of Theobroma cacao. Nature Genetics 43:
101–108.
ARNHEIM, N., M. KRYSTAL, R. SCHMICKEL, G. WILSON, O. RYDER, AND E.
ZIMMER. 1980. Molecular evidence for genetic exchanges among
ribosomal genes on non-homologous chromosomes in man and apes.
Proceedings of the National Academy of Sciences, USA 77: 7323–
7327.
BARDAKCI, F., AND D. O. F. SKIBINSKI. 1994. Application of the RAPD
technique in tilapia fish: Species and subspecies identification.
Heredity 73: 117–123.
BARTLEY, B. G. D. 2005. The genetic diversity of cacao and its utilization.
CABI Publishing, Wallingford, UK.
BELLEMAIN, E., T. CARLSEN, C. BROCHMANN, E. COISSAC, P. TABERLET,
AND H. KAUSERUD. 2010. ITS as an environmental DNA barcode for
fungi: An in silico approach reveals potential PCR biases. BMC
Microbiology 10: 189.
BIRKY, W. C. 2001. The inheritance of genes in mitochondria and chlo-
roplasts: Laws, mechanisms and models. Annual Review of Genetics
35: 125–148.
BLEEKER, W., S. KLAUSMEYER, M. PEINTINGER, AND M. DIENST. 2007.
Chloroplast DNA variations of cultivated radish and its wild relatives.
Plant Science 168: 627–634.
CAMACHO, C., G. COULOURIS, V. AVAGYAN, N. MA, J. PAPADOPOULOS, K.
BEALER, AND T. L. MADDEN. 2009. BLAST+: Architecture and ap-
plications. BMC Bioinformatics 10: 421.
CAVERS, S., C. NAVARRO, AND A. J. LOWE. 2003. Chloroplast DNA phylo-
geography reveals colonization history of a Neotropical tree, Cedrela
odorata L., in Mesoamerica. Molecular Ecology 12: 1451–1460.
CBOL [Consortion of Barcode of Life]. 2009. A DNA barcode for land
plants. Proceedings of the National Academy of Sciences, USA 106:
12794–12797.
CHEESMAN, E. E. 1944. Notes on the nomenclature, classification and
possible relationships of cocoa populations. Tropical Agricure 21:
144–159.
CHOU, C. H., Y. C. CHIANG, AND T. Y. CHIANG. 1999. Within- and between-
individual length heterogeneity of the rDNA-IGS in Miscanthus sinensis
var. glaber (Poaceae): Phylogenetic analyses. Genome 42: 1088–1093.
COART, E., S. VAN GLABEKE, M. DE LOOSE, A. S. LARSEN, AND I. ROLDAN-
HILLIS, D. M., AND S. K. DAVIS. 1988. Ribosomal DNA: Intraspecific
RUIZ. 2006. Chloroplast diversity in the genus Malus: New insights
polymorphism, concerted evolution, and phylogeny reconstruction.
into the relationship between the European wild apple (Malus sylves-
Systematic Zoology 37: 63–66.
tris (L.) Mill.) and the domesticated apple (Malus domestica Borkh).
HILLIS, D. M., C. MORITZ, C. A. PORTER, AND R. J. BAKER. 1991. Evidence
Molecular Ecology 15: 2171–2182.
for biased gene conversion in concerted evolution of ribosomal DNA.
COEN, E., T. STRACHAN, AND G. DOVER. 1982. Dynamics of concerted evo-
Science 251: 308–310.
lution of ribosomal DNA and histone gene families in the melano-
HOWARD, C. 2010. The development of deoxyribonucleic acid (DNA)
gaster species subgroup of Drosophila. Journal of Molecular Biology
based methods for the identification and authentication of medicinal
158: 17–35.
plant material. Ph.D. dissertation, De Montfort University, Leicester,
COOMES, O. T. 2004. Rain forest ‘conservation-through-use’? Chambira
UK. Website http://hdl.handle.net/2086/3972.
palm fibre extraction and handicraft production in a land-constrained
HUSON, D. H., AND D. BRYANT. 2006. Application of phylogenetic net- works
community, Peruvian Amazon. Biodiversity and Conservation 13:
in evolutionary studies. Molecular Biology and Evolution 23: 254–
351–360.
267.
CRONN, R., A. LISTON, M. PARKS, D. S. GERNANDT, R. SHEN, AND T.
IBRAHIM, R. I., J. AZUMA, AND M. SAKAMOTO. 2006. Complete nucleotide
MOCKLER.
sequence of the cotton (Gossypium barbadense L.) chloroplast ge-
2008. Multiplex sequencing of plant chloroplast genomes using
nome with a comparative analysis of sequences among 9 dicot
Solexa sequencing-by-synthesis technology. Nucleic Acids Research
plants. Genes & Genetic Systems 81: 311–321.
36: e122.
INTERNATIONAL COCOA ORGANIZATION. 2011. About cocoa [online].
DEMPEWOLF, H., N. C. KANE, K. L. OSTEVIK, M. GELETA, M. S. BARKER, Z. LAI,
International Cocoa Organization, London, UK. Website http://www.
M. L. STEWART, E. BEKELE, J. M. ENGELS, Q. C. B. CRONK, AND
icco.org/about/growing.aspx [accessed 20 June 2011].
L. H. RIESEBERG. 2010. Establishing genomic tools and resources for
IRISH, B. I., R. GOENAGA, D. ZHANG, R. SCHNELL, S. BROWN, AND J. C.
Guizotia abyssinica (L.f.) Cass.—The development of a library of
MOTAMAYOR. 2010. Microsatellite fingerprinting of the USDA-ARS
expressed sequence tags, microsatellite loci and the sequenc- ing
tropical agriculture research station cacao (Theobroma cacao L.) ger-
of its chloroplast genome. Molecular Ecology Resources 10: 1048–
mplasm collection. Crop Science 50: 656–667.
1058.
IUCN [International Union for Conservation of Nature]. 2006. 2006 IUCN
DEXTER, K. G., T. D. PENNINGTON, AND C. W. CUNNINGHAM. 2010. Using
Red List. Website http://www.iucn.org [accessed December 16, 2011].
DNA to assess errors in tropical tree identifications: How often are
JANSEN, R. K., C. SASKI, S. B. LEE, A. K. HANSEN, AND H. DANIELL. 2011.
ecologists wrong and when does it matter? Ecological Monographs
Complete plastid genome sequences of three rosids (Castanea,
80: 267–286.
Prunus, Theobroma): Evidence for at least two independent trans-
DOORDUIN, L., B. GRAVENDEEL, Y. LAMMERS, Y. ARIYUREK, T. CHIN-A-
fers of rpl22 to the nucleus. Molecular Biology and Evolution 28:
WOENG,
835–847.
AND K. VRIELING. 2011. The complete chloroplast genome of 17 in-
KANE, N. C., AND Q. CRONK. 2008. Botany without borders, barcoding in
dividuals of pest species Jacobaea vulgaris: SNPs, microsatellites
focus. Molecular Ecology 17: 5175–5176.
and barcoding markers for population and phylogenetic studies.
KRESS, W. J., AND D. L. ERICKSON. 2008. DNA-barcoding—A windfall for
DNA Research 18: 93–105.
tropical biology? Biotropica 40: 405–408.
DUGAN, L. E., M. F. WOJCIECHOWSKI, AND L. R. LANDRUM. 2007. A large scale
KRONHOLM, I., O. LOUDET, AND J. DE MEAUX. 2010. Influence of muta-
plant survey: Efficient vouchering with identification through morphol-
tion rate on estimators of genetic differentiation—Lessons from
ogy and DNA analysis. Taxon 56: 1238–1244.
Arabidopsis thaliana. BMC Genetics 11: 33.
EDGAR, R. C. 2004. MUSCLE: Multiple sequence alignment with high ac-
LAHAYE, R. M. van der Bank, D. Bogarin, J. Warner, F. Pupulin, G. Gigot,
curacy and high throughput. Nucleic Acids Research 32: 1792–1797.
O. Maurin, S. Duthoit, T. G. Barraclough, and V. Savalainen. 2008.
ELDER, J. F. JR., AND B. J. TURNER. 1995. Concerted evolution of repetitive
DNA barcoding the floras of biodiversity hotspot. Proceedings of the
DNA sequences in eukaryotes. The Quarterly Review of Biology 70:
National Academy of Sciences, USA 105: 2923–2928.
297–320.
LEE, S.-B., C. KAITTANIS, R. K. JANSEN, J. B. HOSTETLER, L. J. TALLON, C.
FELSENSTEIN, J. 1973. Maximum likelihood and minimum-steps methods for
D. TOWN, AND H. DANIELL. 2006. The complete chloroplast genome
estimating evolutionary trees from data on discrete characters.
sequence of Gossypium hirsutum: Organization and phylogenetic re-
Systematic Zoology 22: 240–249.
lationships to other angiosperms. BMC Genomics 7: 61.
FERRI, G., M. ALU, B. CORRADINI, AND G. BEDUSCHI. 2009. Forensic bot-
Li, H., B. Handsaker, A. Wysoker, T. Fennell, J. Ruan, N. Homer, G.
any: Species identification of botanical trace evidence using a multi-
Marth, G. Abecasis, and R. Durbin, 1000 Genome Project Data
gene barcoding approach. International Journal of Legal Medicine
Processing Subgroup. 2009. The Sequence Alignment/Map
123: 395–401.
(SAM) format and SAMtools. Bioinformatics (Oxford, England)
FIGUEIRA, A., J. JANICK, AND P. GOLDSBROUGH. 1992. Genome size and
25: 2078–2079.
DNA polymorphism in Theobroma cacao. Journal of the American
LOHSE, M., O. DRECHSEL, AND R. BOCK. 2007. OrganellarGenomeDRAW
Society for Horticultural Science 117: 673–677.
(OGDRAW): A tool for the easy generation of high-quality cus-
FITTER, R., AND R. KAPLINSKY. 2001. Who gains from product rents as the
tom graphical maps of plastid and mitochondrial genomes. Current
coffee market becomes more differentiated? A value chain analysis.
Genetics 52: 267–274.
IDS Bulletin 32: 69–82.
LOU, S. K., K. L. WONG, M. LI, P. P. H. BUT, S. K. TSUI, AND P. C. SHAW.
GANLEY, A. R. D., AND T. KOBAYASHI. 2007. Highly efficient concerted
2010. An integrated web medicinal materials DNA database:
evolution in the ribosomal DNA repeats: Total rDNA repeat
MMDBD (Medicinal Materials DNA Barcode Database). BMC
variation revealed by whole-genome shotgun sequence data.
Genomics 11: 402.
Genome Research 17: 184–191.
MEYERS, S., AND A. LISTON. 2010. Characterizing the genome of wild
GLENN, T. C. 2011. Field guide to next-generation DNA sequencers.
relatives of Limnanthes alba (meadowfoam) using massively parallel
Molecular Ecology Resources.
sequencing. Acta Horticulturae 859: 309–314.
GUINDON, S., AND O. GASCUEL. 2003. A simple, fast, and accurate algorithm
MORIN, P. A., J. J. MOORE, AND D. S. WOODRUFF. 1992. Identification of
to estimate large phylogenies by maximum likelihood. Systematic
chimpanzee subspecies with DNA from air and allele specific probes.
Biology 52: 696–704.
Proceedings. Biological Sciences 249: 293–297 .
HEBERT, P. D. N., A. CYWINSKI, S. L. BALL, AND J. R. DEWAARD. 2003.
MOTAMAYOR, J. C., P. LACHNEAUD, J. W. DA SILVA E MOTA, R. LOOR, D. N.
Biological identifications through DNA barcodes. Proceedings.
KUHN, J. S. BROWN, AND R. J. SCHNELL. 2008. Geographic and genetic
Biological Sciences 270: 313–321.
population differentiation of the Amazonian chocolate tree (Theobroma
HILLIER, L. W., G. T. MARTH, A. R. QUINLAN, D. DOOLING, G. FEWELL, D.
cacao L). PLoS ONE 3: e3311.
BARNETT, P. FOX, ET AL. 2008. Whole-genome sequencing and vari- ant
discovery in C. elegans. Nature Methods 5: 183–188.
MOTILAL, L. A., D. ZHANG, P. UMAHARAN, S. MISCHKE, M. BOCCARA, AND
SOLTIS, D. E., M. A. GITZENDANNER, D. D. STRENGE, AND P. S. SOLTIS. 1997.
S. PINNEY. 2009. Increasing accuracy and throughput in large-scale
Chloroplast DNA intraspecific phylogeography of plants from the
microsatellite fingerprinting of cacao field germplasm collections.
Pacific Northwest of North America. Plant Systematics and
Tropical Plant Biology 2: 23–37.
Evolution 206: 353–373.
MUELLNER, A. N., H. GREGER, AND C. M. PANNELL. 2009. Genetic diver-
SOSNICKI, A. A., AND S. NEWMAN. 2010. The support of meat value chains by
sity and geographic structure in Aglaia elaeagnoidea (Meliaceae,
genetic technologies. Meat Science 86: 129–137.
Sapindales), a morphologically complex tree species, near the two
SOLTIS, D. E., A. E. SENTERS, M. J. ZANIS, S. KIM, J. D. THOMPSON, P. S. SOLTIS,
extremes of its distribution. Blumea—Biodiversity. Evolution and
L. P. RONSE DE CRAENE, P. K. ENDRESS, AND J. S. FARRIS. 2003.
Biogeography of Plants 54: 207–216.
Gunnerales are sister to other core eudicots: Implications for the evo-
NEWMASTER, S. G., AND S. RAGUPATHY. 2010. Ethnobotany genomics—
lution of pentamery. American Journal of Botany 90: 461–470.
Discovery and innovation in a new era of exploratory research.
STEELE, P. R., AND J. C. PIRES. 2011. Biodiversity assessment: State-of-the-
Journal of Ethnobiology and Ethnomedicine 6: 2.
art techniques in phylogenomics and species identification.
NEWTON, A. C. 2008. Conservation of tree species through sustainable use:
American Journal of Botany 98: 415–425.
How can it be achieved in practice? Oryx 42: 195–205.
STRAUB, S. C. K., M. FISHBEIN, T. LIVSHULTZ, Z. FOSTER, M. PARKS, K.
NOCK, C. J., D. L. WATERS, M. A. EDWARDS, S. G. BOWEN, N. RICE, G. M.
WEITEMIER, R. C. CRONN, ET AL. 2011. Building a model: Developing
CORDEIRO, AND R. J. HENRY. 2011. Chloroplast genome sequences from
genomic resources for common milkweed (Asclepias syriaca) with
total DNA for plant identification. Plant Biotechnology Journal 9:
low coverage genome sequencing. BMC Genomics 12: 211.
328–333.
STRAUB, S. C. K., M. PARKS, K. WEITEMIER, M. FISHBEIN, R. C. CRONN, AND
OVCHARENKO, I., G. G. LOOTS, R. C. HARDISON, W. MILLER, AND L. STUBBS.
A. LISTON. 2012. Navigating the tip of the genomic iceberg: Next-
2004. zPicture: Dynamic alignment and visualization tool for ana-
generation sequencing for plant systematics. American Journal of
lyzing conservation profiles. Genome Research 14: 472–477.
Botany 99: 349–364.
PALMER, J. D. 1985. Evolution of chloroplast and mitochondrial DNA in
WARD, J., S. R. GILMORE, J. ROBERTSON, AND R. PEAKALL. 2009. A grass
plants and algae. In R. J. MacIntyre [ed.], Monographs in evolution-
molecular identification system for forensic botany: A critical evalu-
ary biology: Molecular evolutionary genetics, 131–240. Plenum,
ation of the strengths and limitations. Journal of Forensic Sciences
New York, New York, USA.
54: 1254–1260.
PARKS, M., R. CRONN, AND A. LISTON. 2009. Increasing phylogenetic reso-
WENDEL, J. F., AND V. A. ALBERT. 1992. Phylogenetics of the cotton genus
lution at low taxonomic levels using massively parallel sequencing
(Gossypium): Character-state weighted parsimony analysis of
of chloroplast genomes. BMC Biology 7: 84.
chloro- plast-DNA restriction site data and its systematic and
PENNISI, E. 2007. Wanted: A barcode for plants. Science 318: 190–191.
biogeographic implications. Systematic Botany 17: 115–143.
PETIT, R. J., AND G. G. VENDRAMIN. 2007. Plant phylogeography based on
Whitlock, M. 2011. G´ST and D do not replace FST. Molecular Ecology
organelle genes: An introduction. In S. Weiss and N. Ferrand [eds.],
20: 1083–1091.
Phylogeography of southern European refugia, 23–97. Springer,
WHITTALL, J. B., J. SYRING, M. PARKS, J. BUENROSTRO, C. DICK, A. LISTON,
Dordrecht, Netherlands.
AND
PIREDDA, R., C. S. MARCO, M. ATTIMONELLI, R. BELLAROSA, AND B.
R. CRONN. 2010. Finding a (pine) needle in a haystack: Chloroplast
SCHIRONE.
genome sequence divergence in rare and widespread pines.
2011. Prospects of barcoding the Italian wild dendroflora: Oaks re-
Molecular Ecology 19 (supplement 1): 100–114.
veal severe limitations to tracking species identity. Molecular
WOLFE, K. H., W. H. LI, AND P. M. SHARP. 1987. Rates of nucleotide
Ecology Resources 11: 72–83.
substitution vary greatly among plant mitochondrial, chloroplast, and
POSADA, D. 2008. jModelTest: Phylogenetic model averaging. Molecular
nuclear DNAs. Proceedings of the National Academy of Sciences,
Biology and Evolution 25: 1253–1256.
USA 84: 9054–9058.
RAGUPATHY, S., S. G. NEWMASTER, M. MURUGESAN, AND V.
WOOD, G. A. R., AND R. A. LASS. 2001. Cocoa, 4th ed. Longman Group,
BALASUBRAMANIAM.
Blackwell, UK.
2009. DNA barcoding discriminates a new cryptic grass species re-
WYMAN, S. K., R. K. JANSEN, AND J. L. BOORE. 2004. Automatic annota-
vealed in an ethnobotany study by the hill tribes of the Western
tion of organellar genomes with DOGMA. Bioinformatics (Oxford,
Ghats in southern India. Molecular Ecology Resources 9: 164–171.
England) 20: 3252–3255.
ROGERS, S. O., AND A. J. BENDICH. 1987. Ribosomal RNA genes in
YANG, J. Y., A. LAMBERT, L. A. MOTILAL, H. DEMPEWOLF, K. MAHARAJ, AND
plants: Variability in copy number and in the intergenic spacer.
Q. C. B. CRONK. 2011. Chloroplast microsatellite primers for cacao
Plant Molecular Biology 9: 509–520.
(Theobroma cacao). American Journal of Botany 98: e372–e374.
ROSS, B. C., K. RAIOS, K. JACKSON, AND B. DWYER. 1992. Molecular clon-
YAO, H., J. SONG, C. LIU, K. LUO, J. HAN, Y. LI, X. PANG, H. XU, Y. ZHU, P.
ing of a highly repeated DNA element from Mycobacterium tuber-
XIAO, AND S. CHEN. 2010. Use of ITS2 region as the universal DNA
culosis and its use as an epidemiological tool. Journal of Clinical
barcode for plants and animals. PLoS ONE 5: e13102.
Microbiology 30: 942–946.
YESSON, C., R. T. BÁRCENAS, H. M. HERNÁNDEZ, M. DE LA LUZ RUÍZ-
ROWNTREE, J. K., R. S. COWAN, M. LEGGETT, M. M. RAMSAY, AND M. F.
MAQUEDA, A. PRADO, V. M. RODRÍGUEZ, AND J. A. HAWKINS. 2011.
FAY. 2010. Which moss is which? Identification of the threatened
DNA barcodes for Mexican Cactaceae, plants under pressure from
moss Orthodontium gracile using molecular and morphological tech-
wild collecting. Molecular Ecology Resources 11:
niques. Conservation Genetics 11: 1033–1042. Schaal, B. A., D. A.
ZHANG, Y.-J., P.-F. MA, AND D.-Z. LI. 2011. High-throughput sequenc- ing
Hayworth, K. M. Olsen, J. T. Rauscher, and W. A. Smith. 1998.
of six bamboo chloroplast genomes: Phylogenetic implications for
Phylogeographic studies in plants: Problems and prospects.
temperate woody bamboos (Poaceae: Bambusoideae). PLoS ONE 6:
Molecular Ecology 7: 465–474.
e20596.
SCHAAL, B. A., AND G. H. LEARN JR. 1988. Ribosomal DNA variation within
ZURAWSKI, G., AND M. T. CLEGG. 1987. Evolution of higher-plant chlo-
and among plant populations. Annals of the Missouri Botanical
roplast DNA-encoded genes: Implications for structure–function
Garden 75: 1207–1216.
and phylogenetic studies. Annual Review of Plant Physiology 38:
SCHWARTZ, S., W. J. KENT, A. SMIT, Z. ZHANG, R. BAERTSCH, R. C.
391–418.
HARDISON,
ZWICKL, D. J. 2006. Genetic algorithm approaches for the phylogenetic
D. HAUSSLER, AND W. MILLER. 2003. Human-mouse alignments with
analysis of large biological sequence datasets under the maximum
BLASTZ. Genome Research 13: 103–107.
likelihood criterion. Ph.D. dissertation, University of Texas at
SIMPSON, J. T., K. WONG, S. D. JACKMAN, J. E. SCHEIN, S. J. JONES, AND I.
Austin, Austin, Texas, USA.
BIROL. 2009. ABySS: A parallel assembler for short read sequence
data. Genome Research 19: 1117–1123.