Jou r
ogy ol
l of Primat na
Journal of Primatology
Bailey et al., J Primatol 2016, 5:1 http://dx.doi.org/10.4172/2167-6801.1000132
ISSN: 2167-6801
Research Article Research Article
Open OpenAccess Access
Evaluating the Genetic Diversity of Three Endangered Lemur Species (Genus: Propithecus) from Northern Madagascar Carolyn A Bailey1, Adam T McLain2, Sébastien Rioux Paquette3, Susie M McGuire4, Gary D Shore1, Runhua Lei1, Jean Claude Randriamanana5, Joseph Désiré Rabekinaja5, Gilbert Rakotoarisoa5, Andriamahery Razafindrakoto6, Rick A Brenneman7, Melissa TR Hawkins1 and Edward E Louis Jr1,5* Bill and Berniece Grewcock Center for Conservation and Research, Omaha’s Henry Doorly Zoo and Aquarium, 3701 South 10th Street, Omaha, NE 68107, USA State University of New York Polytechnic Institute, 100 Seymour Road, Utica, NY 13502, USA 3 University of Sherbrooke, Sherbrooke, QC J1K 2R1, Canada 4 Conservation Fusion, 504 Ridgewood Dr, Bellevue, NE 68005, USA 5 Madagascar Biodiversity Partnership, VO 12 Bis A, Antananarivo 101, Madagascar 6 Department of Animal Biology, University of Antananarivo, Antananarivo 101, Madagascar 7 Giraffe Conservation Foundation, P.O Box 86099, Eros, Windhoek, Namibia 1 2
Abstract The lemur genus Propithecus, the sifakas, is comprised of nine species endemic to the island of Madagascar. Sifakas, like all lemurs, are under intense threat from human activity. Habitat loss from ongoing and rampant deforestation is the primary threat to these animals, with hunting for bushmeat another increasing problem. One side effect of habitat and population loss is a decline of genetic diversity, as populations are reduced in number and isolated from one another in scattered forest fragments. To better understand the loss of genetic diversity in sifaka populations, 209 adult individuals from three species (Propithecus coquereli, P. perrieri, and P. tattersalli) were sampled across 14 sites in northern Madagascar. Fragments of the mitochondrial DNA were sequenced and individuals were genotyped across 65 microsatellite loci isolated from each of the three surveyed species. Heterozygosity modeling was performed on the dataset to account for the possibility of ascertainment bias. Genetic diversity was found to correspond to population size and evidence of population bottlenecks was detected in all three species. Propithecus perrieri, a critically endangered primate with a restricted distribution and small population size, displayed the lowest level of genetic diversity among the three species examined. The situation facing all of these sifakas is dire and P. perrieri in particular is in need of immediate conservation efforts to preserve remaining populations and mitigate the impact of habitat fragmentation.
Keywords: Cross-species amplification; Population genetics; Ascertainment bias; Microsatellites; Lemurs; Propithecus
Abbreviations: IUCN: International Union for Conservation of Nature; BOR: Bora Special Reserve; HIH: Anjiamangirana Classified Forest; MAR: Mariarano Classified Forest; ABK: Ankarafantsika National Park; LAME: Ankavanana Forest; LABE: Analabe Forest; ANAL: Antobiratsy Forest; FIA: Andrafiamena Classified Forest; BIN: Binara Forest; DAR: Matamena Forest; TAT: Bobankora Forest; NDR: Andranotsimaty Forest; NOFY: Mahabenofy Forest; TSIM: Antobinitsimihety Forest; mtDNA: Mitochondrial DNA; D-loop: Displacement loop or control region; DNA PAST: A 2,400 basepair portion of the mitochondrial; MSA: Microsat Analyser; SPECIES: The species to which the genotyped individual belongs; LOCUS_SPECIES: The species from which the genotyped marker was isolated; CROSSSPECIES: A binary variable indicating if the genotype was the result of a cross-species amplification, either ‘yes’ or ‘no’; GLMMs: Generalized linear mixed models; NULL: The estimated frequency of null alleles for the marker in the population of the genotyped individual and NULL2 its squared value; IAM: Infinite Allele Model; TPM: Two-Phase Mutation Model; SMM: Stepwise Mutation Model; PSA: Proportion of Shared Alleles. Introduction The sifakas (genus: Propithecus) are large-bodied (3.0-8.5 kg) diurnal lemurs that are primarily arboreal folivores [1-8]. The nine species of sifaka live in small, female-dominated social groups comprised of one or more reproducing pair(s) of adults and dependent offspring [1,812]. Population growth is slow in sifakas as a result of long generation times, low number of offspring per female (approximately one birth J Primatol ISSN: 2167-6801 JPMT, an open access journal
per female each year) and heavy investment in infant care, making these species particularly susceptible to the pressure of deforestation and hunting [9,10,13]. The total number of extant individuals varies by species, but all nine sifaka species are endangered and threatened by human activity [8]. As recently as 1994 [14] only three species of sifaka were recognized (P. diadema with four subspecies: candidus, diadema, edwardsi, and perrieri; P. tattersalli; and P. verreauxi with four subspecies: coquereli, coronatus, deckenii, and verreauxi), but subsequent revisions [15,16] have elevated all former subspecies to species status [8]. The four species of sifaka inhabiting northern Madagascar are Coquerel’s sifaka (Propithecus coquereli), Tattersall’s sifaka (P. tattersalli), Perrier’s sifaka (P. perrieri), and silky sifaka (P. candidus). For this study, focus was placed on three of these species with the exclusion of P. candidus. The objective was to examine genetic diversity
*Corresponding author: Edward E Louis Jr, Bill and Berniece Grewcock Center for Conservation and Research, Omaha’s Henry Doorly Zoo and Aquarium, 3701 S. 10th Street, Omaha, NE 68107, USA, Tel: +0014027382095; Fax: +0014027330490; E-mail:
[email protected] Received January 22, 2016; Accepted March 04, 2016; Published March 11, 2015 Citation: Bailey CA, McLain AT, Paquette SR, McGuire SM, Shore GD, et al. (2015) Evaluating the Genetic Diversity of Three Endangered Lemur Species (Genus: Propithecus) from Northern Madagascar. J Primatol 5: 132. doi:10.4172/21676801.1000132 Copyright: © 2015 Bailey CA, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Volume 5 • Issue 1 • 1000132
Citation: Bailey CA, McLain AT, Paquette SR, McGuire SM, Shore GD, et al. (2015) Evaluating the Genetic Diversity of Three Endangered Lemur Species (Genus: Propithecus) from Northern Madagascar. J Primatol 5: 132. doi:10.4172/2167-6801.1000132
Page 2 of 13
of three sifaka species with variable population sizes and areas of distribution (Figure 1). Coquerel’s sifaka (Propithecus coquereli; IUCN Red List classification of Endangered) [17] has the largest range of the three investigated species and is distributed in scattered populations across northwestern Madagascar (Figure 1). As reported by the IUCN Red List, the estimated total population is less than 200,000 individuals and is believed to have recently decreased by more than 50% over approximately 50 years, with numbers continuing to decline [17]. Kun-Rodrigues et al. [18] found that the largest extant population of P. coquereli at Ankarafantsika National Park is around 47,000 individuals with population densities varying by location from 5 to 100 individuals/ km2. Tattersall’s sifaka (P. tattersalli; IUCN Red List classification of Critically Endangered) [19] is found across a fragmented range of approximately 245,000 ha near the town of Daraina in northeastern Madagascar, with an estimated remaining population of 6-10,000 animals [20]. Quéméré et al. [21] estimated a slightly higher extant population of 11-26,000 individuals after undergoing an expansive survey, with the total number of individuals likely to exceed 18,000 animals; with variable population densities (between 34 and 90 individuals/km2) making a precise census difficult. Perrier’s sifaka (P. perrieri; IUCN Red List classification of Critically Endangered) [22] occupies a small and fragmented home range across northern Madagascar in the Analamerana and Andrafiamena Massifs with a likely remaining population of 40% in mammals, >25% in fish and >10% in birds [57]. Even when cross-species amplification is successful it is susceptible to drawbacks, including allelic dropout, homoplasy of alleles [58,59] and ascertainment bias [60,61]. Ascertainment bias typically arises in single-species datasets when researchers select a panel of genetic markers based on genotypes from a small number of individuals that do not reflect the genetic composition of the entire population [62]. In multi-species datasets, ascertainment bias can also arise when markers isolated from the source species are cross-amplified in related species [63-65]. In theory, cross-species amplification provides an opportunity to maximize the usefulness of microsatellite markers otherwise used solely for the source species, but ascertainment bias often limits their applicability in multi-species studies since it is impossible to directly compare genetic diversity values among species for a given set of markers if significant ascertainment bias exists [57,66]. Populations undergoing large declines may also be subject to a genetic bottleneck where alleles are lost at a faster rate than the rate at which heterozygosity decreases. A population undergoing a bottleneck is thus temporarily excluded from mutation-drift equilibrium [67]. Therefore, in the event of a bottleneck heterozygote excess will be observed with respect to the anticipated Hardy-Weinberg proportions. This can be a reliable method of measuring recent declines in effective population size [68]. All lemur species across Madagascar are in danger of continued habitat loss due to logging, mining, and unsustainable agricultural practices [69]. Here we used mitochondrial DNA and microsatellite data to compare the genetic diversity of three sifaka species in northern Madagascar. This comparative analysis is helpful in categorizing lemur species for conservation priorities. We hypothesized that the genetic diversity of these three sifaka species would vary due to the current population size, bottleneck history and opportunity for gene flow across habitat patches. Using a variety of statistical methodologies, J Primatol ISSN: 2167-6801 JPMT, an open access journal
we performed analyses that found genetic diversity corresponded to population size of the three species surveyed. Previous research has examined genetic diversity in lemurs [24]. Studies such as Holmes et al. [68] looked at a select number of populations for a single species, while others like Quéméré et al. [21] examined an entire species range utilizing microsatellite markers derived from within the species and from another closely related species but neglected to account for ascertainment bias. This study is the first use of species-specific microsatellite markers to perform cross-species analysis of genetic diversity and test for ascertainment bias in closely related lemur species occupying a fragmented environment.
Methods Sample collection Samples from 209 sifakas were collected during field expeditions in northern Madagascar that occurred between 2000 and 2008. Animals sampled include 82 P. coquereli, 51 P. perrieri, and 76 P. tattersalli. Coquerel’s sifakas were collected from across its range at four sites: Bora Special Reserve (BOR), Anjiamangirana Classified Forest (HIH), Mariarano Classified Forest (MAR), and Ankarafantsika National Park (ANK; Figure 1). Perrier’s sifakas were sampled from Ankavanana Forest (LAME), Analabe Forest (LABE), and Antobiratsy Forest (ANAL) in Analamerana Special Reserve and Andrafiamena Classified Forest (FIA; Figure 1). Collection of samples from Tattersall’s sifakas took place in the Daraina region within six forest localities: Binara Forest (BIN), Matamena Forest (DAR), Bobankora Forest (TAT), Andranotsimaty Forest (NDR), Mahabenofy Forest (NOFY) and Antobinitsimihety Forest (TSIM; Figure 1). Sample sizes per site were restricted by the Ministère de l’Environnement des Eaux et Forêts, Madagascar National Parks, and United States Fish and Wildlife standard regulations. All lemurs investigated in this study were free-ranging individuals safely immobilized with a CO2 Dan-Inject projection rifle (Knoxville, TN) and Pneu-darts™ (Williamsport, PA) containing 10 mg/kg estimated body weight of Telazol (Fort Dodge Animal Health; Overland Park, Kansas). Blood and tissue samples were collected from each sedated animal and immediately stored in room temperature storage buffer [70]. Samples were ultimately transported to the laboratory at Omaha’s Henry Doorly Zoo and Aquarium and stored at -80°C. A HomeAgain® microchip (Schering-Plough Veterinary Corp.; Kenilworth, New Jersey) was placed subcutaneously between the scapulae of each lemur to field-catalog each animal with a unique recognition code to re-identify all captured individuals during any future immobilizations. The GPS location of all immobilized lemurs was recorded and individuals were safely released back at the site of collection. All animals were handled following the American Society of Mammalogists guidelines [71].
DNA extraction and sequencing Genomic DNA was extracted from all individuals using a phenolchloroform isoamyl extraction protocol [72]. For all samples, the variable region of the mitochondrial DNA (mtDNA) displacement loop or control region (D-loop; ~560 bp) was amplified using the primers [73,74] and conditions described in Mayor et al. [16]. An additional 2,400 bp of mtDNA was amplified in four segments as described in Pastorini et al. [75,76] and is hereafter referred to as the PAST fragment. We examined this fragment to provide additional evidence on the state of genetic diversity in these species. Sequence data for the D-loop and PAST fragments were generated from overlapping segments for
Volume 5 • Issue 1 • 1000132
Citation: Bailey CA, McLain AT, Paquette SR, McGuire SM, Shore GD, et al. (2015) Evaluating the Genetic Diversity of Three Endangered Lemur Species (Genus: Propithecus) from Northern Madagascar. J Primatol 5: 132. doi:10.4172/2167-6801.1000132
Page 4 of 13
confirmation. The samples were electrophoresed on a 1.2% agarose gel to verify the PCR product and purified with Exonuclease I and Shrimp Alkaline Phosphatase [77]. The cleaned products were then cycle sequenced using the BigDye® Terminator v1.1 Cycle Sequencing Kit (Thermo Fisher Scientific Inc., Waltham, MA) and the following thermocycler profile conditions: 95°C for 1 min; 34 cycles of 94°C for 30 sec, 50°C for 45 sec, 72°C for 45 sec; 72°C for 10 min. The sequences were analyzed by capillary electrophoresis with an Applied Biosystems Prism 3100 or 3130xl Genetic Analyzer.
Mitochondrial DNA sequence analyses The sequence fragments were aligned to generate a consensus sequence using Sequencher 5.1 (Gene Codes Corporation; Ann Arbor, Michigan), and the consensus sequences were aligned using ClustalX 1.83 [78]. All sequences have been deposited in GenBank, and are available from the referenced accession numbers KU534997KU535205 and KU535234-KU535442. The number of haplotypes, segregating sites and gene diversity (h) were estimated for the control region. Nucleotide diversity (π, or the mean of pairwise sequence differences) and its standard error were calculated uncorrected as in Nei et al. [79] using Arlequin version 3.0 [80]. Portions of the total genetic variation to divergence were assigned either among or within haplotype groupings. This approach incorporated information on the absolute number of differences among haplotypes as well as on haplotype frequencies. The pattern of
sequence evolution was portrayed using a minimum spanning network generated with the program NETWORK version 4.516 (Figure 2) [81] and Arlequin.
Microsatellite data generation A total of 65 microsatellite loci were examined in this study. Loci isolated in each of the three species of sifaka were selected and checked for cross-amplification between the other two species. All 18 loci described by Rakotoarisoa et al. [82] in P. coquereli were used. Nineteen of the 20 markers described in Razafindrakoto et al. [83] for P. tattersalli were amplified with the exclusion of 93HDZ240 as this marker was later determined to be identical to 93HDZ228. Twentyeight novel microsatellite marker loci were developed from P. perrieri following the protocol of Moraga-Amador et al. [84] and modified as described in Razafindrakoto et al. [83] and are presented in Table 1. PCR amplification of each locus was carried out in a 25 µL reaction volume using a MBS thermocycler (Thermo Electron Corporation, Milford MA) with approximately 50 ng of genomic DNA as a template. Final amplification conditions consisted of 12.5 pmol unlabeled reverse primer, 12.5 pmol fluorescently labeled forward primer, 1.5 mM MgCl2, 200 µM each dNTP, and 0.5 units of Taq DNA polymerase (Promega; Madison, WI). The thermal profile for PCR amplification was 95°C for 5 min, followed by 35 cycles of 95°C for 30 sec, a primerspecific annealing temperature for 30 sec (Table 1) [82,83], 72°C for 30 sec, ending with a final extension of 72°C for 10 min. Allele sizes were
Figure 2: A minimum spanning network showing the phylogenetic relationship between the observed mtDNA D-loop haplotypes in three Propithecus species. The number of nucleotide differences is indicated by a dash in their connecting lines. Missing intermediates are indicated by shadowed circles; haplotypes are indicated by different patterns of circles. The partition of circles with different patterns is based on different sites. The size of the circles approximates the number of samples with a particular haplotype.
J Primatol ISSN: 2167-6801 JPMT, an open access journal
Volume 5 • Issue 1 • 1000132
Citation: Bailey CA, McLain AT, Paquette SR, McGuire SM, Shore GD, et al. (2015) Evaluating the Genetic Diversity of Three Endangered Lemur Species (Genus: Propithecus) from Northern Madagascar. J Primatol 5: 132. doi:10.4172/2167-6801.1000132
Page 5 of 13 Locus 94HDZ2 94HDZ17 94HDZ75 94HDZ102 94HDZ111 94HDZ123 94HDZ133 94HDZ134 94HDZ145 94HDZ152 94HDZ154 94HDZ430 94HDZ443 94HDZ470 94HDZ483 94HDZ564 94HDZ565 94HDZ569 94HDZ573 94HDZ610 94HDZ630 94HDZ642 94HDZ646 94HDZ655
Primer Sequence (5’ to 3’) and Dye F: HEXTGT TTC AAA GAG ATT ATT GGG AT R: GTA TTA CAT TTC TAT TGG ACA GTG AG F: FAMTCA AAG CAG CAA AGA CAA TAA A R: TGG ATG GAT GGG TGA AGG F: HEXCCT TCT GTC ATC ATT TTT CCC R: GTT GTG GCT TTG TTT GTT TTG F:
CAG TGA GGA AGA GAA ATG ATA AAA A
FAM
R: AAA CAA ACC CAA GGA AGG AAC F:
HEX
GCC CAC AGA GAA AAA TCC AG
R: TAC AAC CCC AGG AGC CG F: FAMGAT GGA TTT GCC TGC GG R: TGG GTG GTG TGT CTA ATG GG F: FAMGGA CAG GGG CAA GTG GC R: AAT AAA AGA ATG GCA AAG GAC TG F: FAMTAT TCA AGA CTC TCT ATC GCC TG R: GAA GGG GAA GGG AAA TGG F: FAMCGT AAT GGC ACA GAG ATA GAC AA R: CAG GAG TTG GGT TTG GGT A F: FAMGGA AAT GGT CTC AGG GTC AC R: GGA AAT GGT CTC AGG GTC AC F: HEXACA AGG GTG AGC ACA CAT ACA T R: CCT GGA CCT GGT TTA TTT CTT T F: HEXCCT ATT CTT GCG GGC AGT C R: CCT GTT GGT TCT GTC TCT CTG G F: HEXTGA ACC ATT GTA TCC TGC TTT G R: CTT TCC TCC TTT CTT TTG CTT C F: FAMTGT CCA TAA CCC ACA CTT TCT C R: ACT AAA TCT TAC GGA ATC TTC ACT G F: FAMTAA TCA GTG GAA GAA CAG AGG AGA AT R: CAG GAG TTG GGT TTG GGT A F: HEXGAC ACA CAC ACA GGA TTG GG R: GCA CTG GAG GAC AAA TAC ACA T F: HEXGAT CAG GTG CCT CTT AGA ACA TTA C R: CCA CAG TCT GCC TCC TTG AG F: HEXTCA AAG CAC AAA GGA AAG AAC T R: CTC TGA CTC CGA AAC TGG CT F: FAMGCC CAT CTG GTC CAA CTA AG R: AAA TGA ACT GCT GCT CTA AAT AGG F: FAMTTG AGG ACT TTG AGA ACC ACA G R: CAA GCC CAT CAC TCC ACC F: HEXCTT TTT GCC CCA CTC CAG R: ACC CTT GCC TTC ACC ATT C F: FAMGCC GTA AAC ATC CCC GTC R: CGT TCA GTC CTC CAG CCC F: HEXATA GGG ACA TTT GGG GGT AA R: CTT GTT TTT AGA TTT GGG GTC A F:
GAA GCC CAT TCT CAG ACA AA
FAM
R: CAG GAA AGG TGT AAA GCA TTG 94HDZ662
F:
HEX
TGG AAC AGA TGT GAG ATG CC
R: ACA CCC TCC TCC CCA AAG 94HDZ663
F: HEXGAA CTA TCC CTA CCC CAC TCC R: AAA TCT CAC CTA TGC CTG CC
J Primatol ISSN: 2167-6801 JPMT, an open access journal
Annealing Temp. (°C)
Size Range
GenBank Accession No.
(CA)28(TA)2(CA)6(TA)2(CA)4
50
252-262
KU535206
(GT)12
58
219-225
KU535207
(CA)14
58
138-142
KU535208
(CA)19
58
153-163
KU535209
(CA)4GA(CA)13
58
142-156
KU535210
(CA)11
60
194
KU535211
(CA)7CG(CA)16
60
152-168
KU535212
(CA)19TA(CA)13
60
202-208
KU535213
(CA)8AA(CA)2
62
266-284
KU535214
(CA)18
58
206-216
KU535215
(CA)16
60
136-152
KU535216
(GT)7GC(GT)16
60
148-156
KU535217
(CA)15CG(CA)5
50
213-225
KU535218
(CA)4CCTA(CA)15
60
168-182
KU535219
(CA)9AACACTTACATA(CA)15
58
242-260
KU535220
(GT)13
54
204
KU535221
(GT)19
56
176
KU535222
(CA)16
58
102-118
KU535223
(CA)14T(CA)9
58
201-205
KU535224
(CA)15
58
131-145
KU535225
(GT)3GCAT(GT)16
54
167-177
KU535226
(GT)18
58
194-206
KU535227
(CA)10
60
177-179
KU535228
(CA)16
58
225-243
KU535229
(CA)5(CC(CA)2)3CC(CA)10
60
173-185
KU535230
(CA)7CT(CA)6CT(CA)22
60
165-189
KU535231
Repeat Motif
Volume 5 • Issue 1 • 1000132
Citation: Bailey CA, McLain AT, Paquette SR, McGuire SM, Shore GD, et al. (2015) Evaluating the Genetic Diversity of Three Endangered Lemur Species (Genus: Propithecus) from Northern Madagascar. J Primatol 5: 132. doi:10.4172/2167-6801.1000132
Page 6 of 13 94HDZ681
F: FAMTTC AAA AGG CTC AAA AAT ACA AA R: CCA CTA TCG GAA CCC AGG T
94HDZ683
F: FAMGCC ATT CAG ATT CAC ACC AAG R: GGA AGG ATG TCA GAT TAG AGT G
(CA)22
58
123-137
KU535232
(CA)16AA(CA)17
58
176-204
KU535233
Table 1: Primer sequences and fluorescent dye used, repeat motif, optimized annealing temperature, size range (bp) and GenBank accession numbers of 28 novel microsatellite loci discovered in the Propithecus perrieri genome.
determined by separation of the PCR products via POP-4™ capillary buffer electrophoresed on an ABI 3100 or 3130xl Genetic Analyzer. Fragment length genotypes were assigned by GeneScan using GeneScan™ 500XL ROX™ size standard in the GeneMapper software version 4.0. The dataset was analyzed for errors using MICRO-CHECKER [85] and Microsat Analyser (MSA) [86]. Marker independence was tested following a Bonferroni correction for multiple tests in FSTAT [87,88].
Ascertainment bias analyses A modeling framework to quantify the ascertainment bias among the microsatellite genotypes of the three sifaka species was developed. Since species-specific microsatellites were isolated from each of the targeted species and utilized comprehensively across all individuals, we were able to assess the effects of sourced and non-sourced microsatellites on estimates of genetic diversity. The genotypic data were organized so that each successfully amplified single-locus genotype was treated as an observation, leading to a dataset of 10,053 observations. Each observation consisted of a value of heterozygosity (binary variable: 0 or 1) and the following explanatory variables: SPECIES (the species to which the genotyped individual belongs), LOCUS_SPECIES (the species from which the genotyped marker was isolated), and CROSSSPECIES (a binary variable indicating if the genotype was the result of a cross-species amplification, either ‘yes’ or ‘no’). Preparation of the dataset was performed in the R language [89], using functions implemented in the package PopGenKit [90]. Using this dataset, generalized linear mixed models (GLMMs) were fitted with a binomial error distribution and a logit link function. Initially, ten competing models were tested in order to identify the model that best fit the heterozygosity data on the basis of the Akaike Information Criterion. These models represented different combinations of predictors treated as fixed effects and are listed in Supplementary (Table S1). In addition to SPECIES, LOCUS_SPECIES, CROSS-SPECIES and their interactions, the estimated frequency of null alleles for the marker in the population of the genotyped individual (NULL) and its squared value (NULL2) were also included. Because null alleles strongly effect observed heterozygosity and are more likely to occur in cross-species amplifications due to mutations in microsatellite flanking regions, the inclusion of these variables was required to model heterozygosity while controlling for the presence of null alleles in the dataset. The frequency of null alleles was estimated with the maximumlikelihood procedure implemented in the software ML-Null [91]. All competing models included individual ID, population ID, and marker ID as random factors to account for non-independence of observations (i.e. heterozygosity may vary among individuals, populations, and markers), and the fixed effects LOCUS_SPECIES and CROSS-SPECIES were never included in the same model to avoid redundancy. GLMMs were computed with the R package lme4 [92]. Preliminary results revealed that the model that best explained heterozygosity included the predictors SPECIES, CROSS-SPECIES, the interaction of these predictors, NULL and NULL2. In this model, the value of the crossspecies parameter thus provides a quantitative estimate of the bias in heterozygosity due to the use of microsatellites from a different species. The effect of cross-species amplification was also investigated, as J Primatol ISSN: 2167-6801 JPMT, an open access journal
well as how this varied when using only a subset of the 65 markers. Since only 33 of the 65 markers reliably produced PCR products across all three species, a reasonable approach would be to discard the other 32 before calculating genetic diversity indices (e.g. heterozygosity) to allow comparison among species. Nonetheless, this may also affect the results with respect to those obtained from all available genetic data. Thus, a new binomial GLMM was fitted with only the data from markers that worked in all species. Finally, an objective marker selection process was also performed like the one carried out in all population/conservation genetic studies to retain only markers satisfying a set of criteria, the most limiting of which was a low null allele frequency (
