DNA Barcode of the Flora of Calakmul, Yucatan Peninsula, Mexico

DNA Barcode of the Flora
of Calakmul, Yucatan
Peninsula, Mexico
Lidia I. Cabrera, Gerardo A.
Salazar, Esteban Martínez, Clara
Ramos, Patricia Escalante,
Fernando Chiang, Mario Sousa,
María Kuzmina
1) licabreram@yahoo.com.mx;
2) lcabrera@ibunam2.ibiiologia.unam.mx

Two approaches to barcoding:
• Taxonomic Group
– Expert on a specific taxon
– Large number of samples of the group
• Flora (or biota) from a place with
previous botanic knowledge
• Herbarium samples often available and
already identified by taxonomist

Calakmul: part of the Mayan tropical
forest, second largest rain forest in
the Americas after the Amazon
Calakmul Biosphere
Reserve
•23 398 Km 2
•Limestone substrate
9 types of vegetation
Martinez et al. (2001): 147
families of land plants,
1537 species (70% of the
flora of Campeche).
Imagen de
satélite
I
II

Aims of this study:
• Assess the potential to barcode a tropical flora
from herbarium specimens.
• Determine the performance of the standard
barcode for land plants (rbcL + matK) for the
taxonomic discrimination at family, genus and
species levels.

Material and methods
950 herbariums samples from the project Flora de Calakmul
(MEXU); 2 samples per species (81 families, 301 genera, 566
species)
Geo-referenced
database,
digital pictures and
traces uploaded to
BOLD
(www.boldsystems.org)
Fragments of 5 5
mm (leaf, bract,
tepal)
standard
sequencing
protocols at
Guelph, Canada
Data Analysis:
BLAST in GenBank.
Distance Analysis (NJ, K2P model, for pairwise
genetic distances and the threshold for
distances for discrimination with the program
TaxonDNA).
Parsimony Analysis (bootstrap, 1000 replicates
to assess group support).

Project Specimens
CALAK
DNA
barcode of
the flora
of
Calakmul,
Sequence
quality
stats
Sequencing Results
Species Species
with
sequences
matK
950 566 367
(64.8%)
Species
with
sequences
rbcLa
449
(79.3%)
Species
with
barcodes(matKrbcLa)
301
(53.65%)
Sequences
matK
531
(55.9%)
High
99.85%
Sequences
rbcLa
675
(71.1%)
High
99.62%
Barcodes(matKrbcLa)
403
(42.4%)
(276-933 bp) (390-552 bp)

Taxonomic assignment by
GenBank BLAST
Correct
taxonomic
match in
GenBank
for matK
Correct
txonomic
match in
GenBank
for rbcL
Taxa
correctly
assigned by
matK
Taxa
correctly
assigned by
rbcL
Family level 503 (94.73%) 672 (99.56%) 73 (90.12%) 52 (72.84%)
Genus level 357 (67.23%) 538 (79.70%) 137 (45.51.%) 187 (62.13%)
Species level 98 (18.46%) 147 (21.8%)
68 (12.12%)
89 (15.87%)

matK
531 sequences
rbcL
675 sequences
matK + rbcL
403 sequences
Neighbor joining trees
Trees congruent with
angiosperm
phylogeny (APG III)
Parsimony bootstrap recovers same groups
Family (Bootstrap support)
Acanthaceae (100) Anacardiaceae (100)
Agavaceae (100) Apocynaceae (100)
Araceae (100) Arecaceae (100)
Bignoniaceae (61) Boraginaceae (100)
Bromeliaceae (100) Cactaceae (100)
Canellaceae (100) Cannabaceae (100)
Convolvulaceae (100) Dilleniaceae (100)
Ebenaceae (100) Elaeocarpaceae (100)
Erythroxylaceae (100) Fabaceae (100)
Hypocrateaceae (100) Lauraceae (100)
Malpighiaceace (100) Meliaceae (99)
Moraceae (100) Myrtaceae (100)
Olacaceae (100) Polygalaceae (100)
Polygonaceae (100) Rhamnaceae (91)
Rubiaceae (100) Rutaceae (100)
Sapindaceae (100) Sapotaceae (100)
Solanaceae (100) Sterculiaceae (100)
Violaceae (100) Vitaceae (100)

Distance summary for 403 complete standard
barcodes (matK-rbcL)
400
350
300
250
200
150
100
50
0
Histogram of the number of intraspecific and interspecific comparisons
Series1 Interspecific
Series2 Intraspecific
Overlapping rank between
intra– and interspecific
distances: 0%-0.76%
1.68% “ Best Close Match”
threshold
34.73% of species can be
distinguished with this
threshold

General considerations
• High quality sequences of rbcL and matK could be recovered for 71%
and 56% of the sampled herbarium specimens, respectively.
• Complete standard barcodes (rbcL + matK) were obtained for 42% of
the herbarium specimens using standard protocols (i.e. without
additional optimization effort for individual samples).
• Taxonomic assignment by the BLAST (GenBank) was high at family
level, lower at genus level and lowest at the specific level. This is to be
expected because the greater inclusivity of higher ranks; however, few
Mexican tropical plants have been sequenced for these markers so
there is little chance of a specific match.
• Overlapping between intra-and interspecific distances: 0%-0.76%. The
1.68% threshold (“Best Close Math”) allows distinction of 35% of the
species can distinguish with this.

General considerations
• PCR and sequencing optimization would increase the yield of highquality
standard barcodes from herbarium specimens; although
impractical from a high-throughput barcoding strategy, this would
maximize the potential of museum (herbarium) collections to advance
in creating reference sequence databases for many tropical areas where
sample collection and identification are expensive and timeconsuming.
• Increased sequencing success would increase intraspecific sampling,
making discrimination thresholds more robust (e.g. for “best close
match”).
• Nevertheless, distinct thresholds seem not to exist for the standard
barcode markers for particular groups (e.g. Ficus, Moraceae). In these
cases additional, more variable markers will be necessary for species
distinction.

General considerations
• Herbarium-based barcoding represents a reasonable compromise
between minimizing financial needs (e.g. for collecting expedition an
time consuming, traditional identification) and increased sequencing
effort (PCR and sequencing optimization).
• This project is an example of the great potential of well-curated
herbarium collections to advance in the generation of reference
sequence databases to speed-up floristic inventory in tropical areas,
especially when funding is an issue.

Other projects in process
• Flora of the Tropical Biologic Station “Los
Tuxtlas”, Veracruz (IBUNAM).
• Flora of the Station “Chamela”, Jalisco
(IBUNAM).
• Flora of the dry forest of Zimatán, Oaxaca
(SERBO/IBUNAM).
• Flora of the Pedregal de San Ángel, México, D.F.
(IBUNAM).
• Some Families in Mexico (Araceae, Orchidaceae,
Apocynaceae/Asclepiadoideae).

Acknowledgements
• Mexican DNA Barcoding
Network (MexBOL), National
Council of Science and
Technology (CONACyT).
• Biodiversity Institute of
Ontario, University of Guelph,
Canada.
• Patricia Rosas
• Daniel Hernandez
• Victor Trejo
• Karen Beatriz Hernandez