Firefly Chan Kin Onn

bioRxiv preprint first posted online May. 9, 2019; doi: http://dx.doi.org/10.1101/632612.
The copyright holder for this preprint (which

was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
It is made available under a CC-BY-NC-ND 4.0 International license.
1 Title: Phylogeography and Molecular Species Delimitation Reveal Cryptic and

2 Incipient Speciation in Synchronous Flashing Fireflies (Coleoptera: Lampyridae) of
3 Southeast Asia
4
5 Short running title: Cryptic and Incipient Speciation in Synchronous Flashing Fireflies
6
7 Authors: Wan F. A. Jusoh1, Lesley Ballantyne2 & Chan Kin Onn1*

8
1
9 Lee Kong Chian Natural History Museum, Faculty of Science, National University of
10 Singapore, 2 Conservatory Drive, Singapore 117377.
2
11 School of Agricultural and Wine Sciences, Charles Sturt University, Wagga Wagga
12 2678, Australia
13
14 *Corresponding author: chankinonn@gmail.com

15
16 Keywords: Lampyridae; CO1; CAD; phylogenetics; biogeography; species

17 boundaries; systematics
18
1
bioRxiv preprint first posted online May. 9, 2019; doi: http://dx.doi.org/10.1101/632612. The copyright holder for this preprint (which
19 ABSTRACT
20 Synchronous flashing fireflies of the genus Pteroptyx are ubiquitous throughout
21 Southeast Asia, yet, knowledge on its biodiversity and evolutionary history remains
22 lacking. Recent studies have revealed notable population-level phylogeographic
23 structure within the P. tener and P. bearni groups in Malaysia, suggesting that cryptic
24 species may exist. Additionally, the close morphological and genetic affinity of the
25 recently described species P. balingiana to P. malaccae has raised questions about
26 its validity. In this study, we assembled the most densely sampled genetic dataset on
27 Pteroptyx to-date to estimate a comprehensive phylogeny using mitochondrial and
28 nuclear DNA and subsequently implemented a suite of distance-, phylogeny-, and
29 coalescent-based species delimitation methods to characterize species boundaries
30 within the P. tener, P. bearni, and P. balingiana/P. malaccae groups. Using a total
31 evidence approach from multiple lines of evidence, we showed that populations of P.
32 tener along the west coast of Peninsular Malaysia are sufficiently divergent from
33 populations from the east coast and Borneo to warrant specific recognition, despite
34 the absence of morphological differentiation. Conversely, divergence of P. bearni
35 from Borneo and eastern Peninsular Malaysia, as well as P. balingiana from P.
36 malaccae were modest and their distinction as separate species were ambiguous;
37 consistent with incipient species in the gray zone of speciation. Overall, this study
38 contributes to the limited but growing body of genetic work on Southeast Asian
39 fireflies and underscores the urgent need to increase the breadth and depth of
40 geographic, taxonomic, and genetic sampling to provide a deeper understanding of
41 their biodiversity and evolutionary history.
42
43
2
44 INTRODUCTION
45 Fireflies or lightning bugs are soft-bodied beetles of the family Lampyridae with the
46 number of estimated species ranging from at least 2,000 to possibly 8,000 globally
47 (Lloyd, 2008). Approximately 1,200 firefly species are known from tropical America
48 (Faust, 2004), more than 400 species from Southeast Asia and the Indo-Pacific
49 region (largely from the subfamily Luciolinae; Ballantyne, Lambkin, Boontop, & Jusoh,
50 2015; McDermott, 1966), while little information is available from tropical Africa (Lloyd,
51 2008).
52 Fireflies are best known for the ability of males to emit precise flashing
53 patterns to attract females, although not all species are luminescent (Lloyd, 2002).
54 The most spectacular of all flashing displays is the near-to-perfect synchronous
55 flashing of fireflies from the genus Pteroptyx Olivier (subfamily Luciolinae), where
56 species are known to occur in multitudes on trees and shrubs along tidal rivers of
57 mangrove swamps (Jusoh, Ballantyne, Lambkin, Hashim, & Wahlberg, 2018). Its
58 highly precise synchronous flashing and tree-swarming behaviour have attracted
59 scientific as well public interests to explicate the mechanisms for their synchronicity
60 (Buck, 2004), elucidate congregating and courting behaviour (Case, 2007; Lloyd,
61 Wing, & Hongtrakul, 2006) and recently, develop as icon species for sustainable
62 ecotourism and conservation (Khoo et al. 2012). However, although fireflies have
63 been the focus of numerous ecological, physiological, and even biomedical research
64 (Copeland & Moiseff, 2006; Ermentrout, 1991; Fraga, 2008; Gould Stephen J. &
65 Subramani Suresh, 1988; Schena, Griss, & Johnsson, 2015), fundamental
66 knowledge on its biodiversity and evolutionary history is still lacking and underscores
67 the dire need for more intensive research.
68 According to recent studies (Ballantyne et al., 2015; Jusoh et al., 2018),
69 Pteroptyx is an Oriental genus comprising 18 species distributed from Hong Kong,
70 southwards throughout Southeast Asia including India (Madras) (Ballantyne, Fu, Shih,
71 Cheng, & Yiu, 2011; Ballantyne & McLean, 1970) and the Philippines (Ballantyne,
72 2001). Most species in this genus are known to occupy mangroves, riparian, or
73 coastal habitats (Ballantyne et al., 2011; Jusoh et al., 2018; Jusoh, Hashim, &
74 Ibrahim, 2010). For instance, at least eight out of 12 species of Pteroptyx were
75 recorded from mangroves throughout Malaysia (Jusoh et al., 2018), followed by four
76 species in Thailand (Sartsanga, Swatdipong, & Sriboonlert, 2018), two to three
3
77 species in Singapore (Chan, Ballantyne, & Goh, 2012), and one species in the
78 Phillippines (Ballantyne et al., 2011; Jusoh et al., 2018). For centuries, species
79 descriptions in the subfamily Luciolinae, which includes Pteroptyx, have been based
80 purely on morphology. However, without morphological characters of the male
81 genitalia and highly trained taxonomists, fireflies are difficult to identify to the species
82 level (Ballantyne, 2012). Moreover, it is almost impossible to provide accurate
83 identification of female and larval specimens, even to the genus level unless the
84 specimens were collected in association with the male (Jusoh et al., 2018). Thus,
85 additional methods and data streams are needed to facilitate species discovery in
86 order to realize the true extent of firefly diversity in the region.
87 Recent developments in molecular sequencing technology have greatly
88 improved our ability to detect, discover, and describe new species (Sites, Marshall, Jr,
89 & Marshall, 2003; Wiens, 2007). Furthermore, comparing profiles of genetic
90 divergences from informative markers have been shown to be an effective way to
91 identify species and screen for taxonomic incompatibilities and potential new species
92 (Chan & Grismer, 2019; Chan, Grismer, & Brown, 2018; Fouquet et al., 2007; Hebert,
93 Cywinska, Ball, & DeWaard, 2003; Padial & De La Riva, 2007; Vences, Thomas,
94 Bonett, & Vieites, 2005; Vences, Thomas, Meijden, Chiari, & Vieites, 2005; Vieites et
95 al., 2009). Among the genetic markers implemented for these purposes, the DNA
96 barcode based on the CO1 mitochondrial gene is one of the most widely used,
97 especially among invertebrates. Although it was initially developed for species
98 identification (Hebert et al., 2003), this robust marker has also been used effectively
99 for species discovery and delimitation (Cao et al., 2016; Hebert & Gregory, 2005;
100 Hubert & Hanner, 2016; Sheth & Thaker, 2017; Yang & Rannala, 2017).
101 DNA sequence data for fireflies, in particular the CO1 gene, have recently
102 been accumulating but a well-characterized profile of genetic variation and a justified
103 threshold for species boundaries have yet to be determined. The first study that
104 implemented DNA barcoding on several species of Pteroptyx in Malaysia found
105 congruence between morphology and molecular data (Jusoh, Hashim, Sääksjärvi,
106 Adam, & Wahlberg, 2014). However, molecular results also revealed considerable
107 phylogeographic structure (Fig. 2 in Jusoh et al., 2014), suggesting that the genus
108 Pteroptyx could potentially harbour cryptic or undiscovered species. In particular,
109 Pteroptyx tener Olivier was clustered into three geographically isolated clades: along
4
110 the east and west coast of Peninsular Malaysia (PM east and PM west), and Borneo
111 (Fig. 1). In Thailand, a population of P. tener was discovered for the first time in 2015,
112 which prompted the hypothesis that geographic polymorphisms may exist
113 (Sriboonlert, Swatdipong, Wonnapinij, E-Kobon, & Thancharoen, 2015). Subsequent
114 examination of morphological characters did not reveal differences that were
115 congruent with the geographically circumscribed haplotypes (Jusoh et al., 2018).
116 Pteroptyx bearni Olivier on the other hand, exhibited subtle differences in colouration
117 and was geographically and genetically clustered into two clades: PM east and
118 Borneo (Jusoh et al., 2014).
119 Pteroptyx malaccae (Gorham, 1880) is a widely distributed species in
120 mangroves throughout Southeast Asia (Ballantyne, 1987; Ballantyne & McLean,
121 1970). In Malaysia, P. malaccae occurs in small populations, usually in sympatry with
122 P. tener, which appears in much bigger congregations (Foo & Dawood, 2017; Jusoh
123 et al., 2018). Meanwhile in Thailand, P. malaccae forms mass congregations in trees
124 along the riverbanks in sympatry with Pteroptyx valida Olivier (Prasertkul, 2018;
125 Sartsanga et al., 2018). Jusoh et al. (2018) hypothesized that P. malaccae is a
126 morphologically variable species complex that requires further investigation after four
127 morphologically distinct groups (Ballantyne, 2001) were detected using a
128 combination of morphological and genetic data. The same analysis, although less
129 resolved, also revealed that P. malaccae formed a sister relationship with a newly
130 described species, Pteroptyx balingiana Jusoh. The authors previously argued that
131 the latter species (although misidentified as Pteroptyx gelasina Ballantyne, see Jusoh
132 et al. 2018) could be “isolated from the P. malaccae by distinct geographically
133 structured variation in the DNA barcodes” (Jusoh et al., 2014).
134 Both previous studies (Jusoh et al., 2018, 2014) suggested that the spatio-
135 genetic structure of populations within Pteroptyx tener, P. bearni, and P. malaccae/P.
136 balingiana could represent complexes containing cryptic and undescribed species,
137 but no explicit species delimitation analyses have been performed to validate those
138 hypotheses. In this study, we assembled the most densely sampled genetic dataset
139 of Pteroptyx to-date to estimate a comprehensive phylogeny using mitochondrial and
140 nuclear DNA. We then employed a suite of distance-, phylogeny-, and coalescent-
141 based species delimitation methods to test whether 1) geographically structured
142 populations in P. tener and P. bearni represent cryptic species; and 2) Pteroptyx
5
143 balingiana is sufficiently genetically distinct from P. malaccae to warrant specific

144 recognition.
145
146 METHODS
147 Taxon coverage and geographic sampling
148 We used sequence data from published studies that are publicly available on
149 GenBank and constructed a dataset consisting of 157 sequences of the
150 mitochondrial Cytochrome c oxidase subunit 1 (CO1) gene and 39 sequences of the
151 nuclear protein coding gene, carbamoylphosphate synthetase (CAD; Supplementary
152 information, Table S1). Apart from the focal taxa (Pteroptyx malaccae, P. balingiana,
153 P. tener, and P. bearni), we also included other closely related taxa such as
154 Pteroptyx asymmetria Ballantyne and P. valida, while outgroup samples include
155 Pteroptyx galbina Jusoh, Pteroptyx testacea (Motschulsky), Colophotia praeusta
156 (Eschscholtz) and Colophotia brevis Olivier. The focal taxa were sampled from 23
157 unique localities across Thailand and Malaysia (Fig. 1). The CO1 gene comprises
158 three separate fragments, which include the prominent “Folmer region” of a 665 bp
159 fragment located at the 5’ end of the CO1 gene (Folmer, Black, Hoeh, Lutz, &
160 Vrijenhoek, 1994), a ~540 bp fragment at the 3’ end (Villalba, Lobo, Martin-Piera, &
161 Zardoya, 2002), and a long fragment (~1,310 bp) that overlaps with the first two
162 fragments (Sartsanga et al., 2018). Samples that contain more than one of these
163 fragments were combined to form a single contig. Sequences from the second CO1
164 fragment which were amplified using the primers CO1-Sca (F)/CO1-Sca (R) (Villalba
165 et al., 2002) overlapped with the Cytochrome c oxidase subunit 2 (CO2) gene as well
166 as a tRNA-Leu and was removed prior to sequence alignment. Both CO1 and CAD
167 sequences were checked for stop codons and aligned separately using MUSCLE
168 with default parameters performed on MEGA-X version 10.0.5 (Kumar, Stecher, Li,
169 Knyaz, & Tamura, 2018). Alignment files are available as Supporting information.
170
171 Phylogenetic analyses
172 We used the program IQ-TREE v1.6 (Nguyen, Schmidt, Von Haeseler, & Minh, 2015)
173 to estimate a maximum likelihood (ML) phylogeny, while a Bayesian phylogeny was
6
174 inferred using BEAST v2.5 (Bouckaert et al., 2014). For the ML phylogeny,
175 sequences were partitioned by gene and the best-fit model of DNA evolution for each
176 partition was determined using ModelFinder (Kalyaanamoorthy, Minh, Wong, Von
177 Haeseler, & Jermiin, 2017). Branch support was assessed with 5,000 bootstrap
178 replicates using Ultrafast Bootstrap Approximation (UFB; Hoang et al. 2017). UFB
179 values above 95 were considered well supported. The BEAST analysis was
180 implemented through the CIPRES portal (Miller, Pfeiffer, & Schwartz, 2010) and the
181 best-fit substitution model for each partition was estimated via model averaging using
182 the bModelTest plugin in BEAST (Bouckaert & Drummond, 2017). A relaxed log-
183 normal and Yule model was used as the molecular clock and tree priors respectively,
184 while all other priors were set to default values. We executed two separate MCMC
185 chains at 50 million generations each and checked for convergence using the
186 program Tracer v1.6 (Rambaut, Drummond, Xie, Baele, & Suchard, 2018). Sampled
187 trees from both MCMC runs were combined using logcombiner and a maximum
188 clade credibility tree was constructed using treeannotator with a burn-in of 10% and
189 nodes summarized using mean heights.
190
191 Species delimitation
192 Genetic distance and ABGD.—we calculated uncorrected p-distances within and
193 between taxa using MEGA-X (Kumar et al., 2018) to obtain a profile of intra- and
194 interspecies divergence thresholds. Subsequently, the Automatic Barcoding Gap
195 Discovery (ABGD) method was used to detect breaks between intraspecific and
196 interspecific diversity known as the barcode gap (Puillandre, Lambert, Brouillet, &
197 Achaz, 2012). The ABGD uses pairwise distances to determine divergence between
198 sequences and does not require a priori specification of divergence thresholds.
199 Because ABGD analyzes single-locus data and has been shown to be effective with
200 the CO1 gene, we performed this analysis on the CO1 alignment. The analysis was
201 performed through the web-server
202 (http://wwwabi.snv.jussieu.fr/public/abgd/abgdweb.html) using default settings and
203 the Kimura 2-parameter (K80) distance model (Ratnasingham & Hebert, 2013).
204 bGMYC.—compared to ABGD, the Bayesian implementation of the general mixed

205 Yule-coalescent (bGMYC) is a non-distance-based method that does not rely on
7
206 similarity threshold parameters. It models the Yule and coalescent processes on an
207 ultrametric tree to determine the transition between intra and interspecific
208 divergences. To account for potential errors in phylogenetic estimation and
209 uncertainty in model parameters, this method integrates over uncertainty in tree
210 topology and branch lengths via MCMC (Reid & Carstens, 2012). As input, we used
211 100 randomly selected trees that were sampled from the posterior distribution of the
212 BEAST analysis. For each tree, we ran the MCMC sampler for 50,000 generations
213 with a burnin of 40,000, retaining 10,000 post burn-in generations with a thinning
214 interval of 100. We then assessed a range of probability thresholds ranging from
215 conservative (0.05) to moderate (0.25) and liberal (0.5) delimitation schemes. The
216 analysis outputs results in the form of posterior probabilities that sequences are
217 conspecific. Hence, we consider populations with PP<0.05 as strong support, 0.1<
218 PP< 0.05 as moderate support, and PP > 0.1 as weak support for splitting.
219 mPTP.—the multirate Poisson tree process (mPTP) is also a phylogeny-aware

220 method that also does not rely on similarity threshold parameters. Similar to bGMYC,
221 it models the transition between intra- and interspecific divergence that is determined
222 by the coalescent and speciation parameters, where intraspecific branching events
223 are expected to be more frequent compared to among species (Kapli et al., 2017;
224 Zhang, Kapli, Pavlidis, & Stamatakis, 2013). However, mPTP differs from bGMYC in
225 modelling speciation and coalescent events relative to numbers of substitutions
226 instead of time (Tang, Humphreys, Fontaneto, & Barraclough, 2014). It takes into
227 account the evolutionary relationships of sequences and uses the number of
228 accumulated expected substitutions between subsequent speciation events to model
229 the branching process. We used the ML phylogeny as the input tree and confidence
230 of delimitation schemes were assessed using two independent MCMC chains at
231 5,000,000 generations each. Support values represent the fraction of sampled
232 delimitations in which a node was part of the speciation process.
233 BPP.—the BPP analysis uses a Bayesian modelling approach to estimate posterior
234 probabilities of species assignments using gene trees, while taking into account
235 uncertainties in the coalescent process (Yang & Rannala, 2010). The A10 analysis
236 (species delimitation using a fixed guide tree) was implemented using relationships
237 derived from phylogenetic analyses as a guide tree. We used a diffuse prior of α =3
238 for both θ and τ priors and the corresponding β parameter was adjusted according to
8
239 the mean (m) estimate of nucleotide diversity (for θ) and node height (for τ ) using the
240 equation m=β/(α-1), for α>2 (Flouri, Jiao, Rannala, & Yang, 2018). The mean root
241 node height was obtained from the BEAST phylogeny. The use of empirically-
242 estimated priors has been shown to provide more reliable results compared to priors
243 selected using non-objective methods (Chan & Grismer, 2019). To accommodate
244 uncertainty in the guide tree, we also performed the A11 analysis (joint species
245 delimitation and species-tree estimation). The MCMC was set to 100,000 samples
246 with burnin=10,000 and sample frequency=5. Convergence was assessed by
247 comparing the consistency of posterior distributions (Flouri et al., 2018; Leaché, Zhu,
248 Rannala, & Yang, 2018; Ziheng, 2015). Posterior probabilities, PP ≥ 0.95 were
249 considered highly supported, 0.90 ≤ PP < 0.95 were considered moderately
250 supported, whereas PP < 0.9 were considered weakly supported.
251 Heuristic gdi.—because Bayesian model selection in BPP can sometimes oversplit
252 species by detecting population splits as species divergence (Leaché et al., 2018;
253 Sukumaran & Knowles, 2017), we used the heuristic genealogical index index (gdi)
254 that has been shown to produce more accurate results (Chan & Grismer, 2019;
255 Jackson, Carstens, Morales, & O’Meara, 2017a; Leaché et al., 2018). First, the A00
256 analysis in BPP was implemented to generate posterior distributions for the
257 parameters τ and θ using the same empirically-derived priors. Four separate runs
258 were performed to ensure convergence and converged runs were combined to
259 generate posterior distributions for the MSC parameters that were subsequently used
260 to calculate the gdi following the equation: gdi = 1 – e-2τ/ θ (Jackson, Carstens,
261 Morales, & O’Meara, 2017b; Leaché et al., 2018). Population A is distinguished from
262 population B by using 2τAB/θA, while 2τAB/θB is used to differentiate population B from
263 population A. Populations are considered distinct species when gdi values are >0.7,
264 while low gdi values <0.2 indicate that populations belong to the same species.
265 Values of 0.2> gdi < 0.7 indicate ambiguous species status (Jackson et al., 2017b;
266 Pinho & Hey, 2010).
267
268 RESULTS
269 Phylogenetic analyses
9
270 Both ML and Bayesian phylogenies produced congruent topologies with high support
271 across most major nodes (Figs. 1, S1, S2). Known species also formed monophyletic
272 clades with high support. Pteroptyx valida, P. asymmetria, and P. malaccae did not
273 show geographically circumscribed genetic sub-structuring, with rampant mixing
274 occurring between individuals from Peninsular Malaysia (PM) and Thailand.
275 Pteroptyx balingiana was reciprocally monophyletic with P. malaccae with high
276 support, whereas the P. bearni and P. tener clades contained multiple populations
277 that were distinctly structured according to geography. Pteroptyx bearni formed two
278 distinct clades (Borneo and PM), while P. tener was structured into three relatively
279 divergent clades. The Bornean population of P. tener formed a sister relationship with
280 populations from eastern PM and this clade (Borneo + PM east) was in turn
281 reciprocally monophyletic with populations from western PM (Fig. 1). Using this
282 phylogeographic framework, we defined the following geographically circumscribed
283 and reciprocally monophyletic population pairs as candidate species for downstream
284 species delimitation analyses: i) balingiana vs. malaccae; ii) bearni Borneo vs. bearni
285 PM; iii) tener Borneo vs. tener PM east; and iv) tener PM west vs. tener Borneo+PM
286 east (Fig. 1).
287
288 Genetic divergence and species delimitation
289 Genetic distance.—the interspecies genetic divergence between the sister lineages P.
290 bearni/ P. tener and P. valida/P. malaccae were high (6.3–8.9% and 8–10.5%
291 respectively), while divergences among populations of the same species were low
292 (<3%). For the candidate species, divergences were slightly higher compared to
293 intrapopulation distributions, except for tener PM west vs. tener Borneo + tener PM
294 east which was higher (3.15–5.49%), but still well below interspecific divergences
295 (Fig. 2).
296 ABGD.—a total of seven partitions were inferred with prior maximal
297 intraspecific distances (P) ranging from 0.001 to 0.02 (Fig. 3; Table 1). For the initial
298 run, the number of delimited species rapidly plateaued at 10 species. This
299 delimitation scheme lumped balingiana with malaccae, bearni Borneo with bearni PM,
300 and tener Borneo with tener PM east. However, tener PM west was delimited as a
301 distinct species from tener Borneo + tener PM east (Table 2). At P=0.0077, the
10
302 recursive run also recovered the same 10 species but at P=0.0046 (partition 4), tener
303 Borneo was further split from tener PM east. Partitions 1–3 of the recursive run
304 produced 13–26 species, which we consider as erroneous as some of the delimited
305 groups were not monophyletic.
306 bGMYC.—the distribution of the coalescence to Yule branching rate ratios

307 were well above zero, indicating that the model is a good fit to the data. The results
308 showed low support for the splitting of balingiana/malaccae (conspecificity PP=0.29),
309 bearni Borneo/PM (PP=0.23), and tener Borneo/PM east (PP=0.5). On the other
310 hand, the splitting of tener Borneo + tener PM east/tener PM west was moderately
311 supported (PP=0.08; Table 2). At the most conservative threshold of 0.05, the
312 analysis lumped balingiana with malaccae, and all populations of P. bearni and P.
313 tener. At 0.25, balingiana remained lumped with malaccae, while tener Borneo was
314 still lumped with tener PM east. However, bearni Borneo was split from bearni PM
315 and tener PM west was split from tener Borneo + tener PM east. The most liberal
316 threshold of 0.5 further split balingiana from malaccae and tener Borneo from tener
317 PM east. Certain samples of P. malaccae from Thailand were also split from other P.
318 malaccae (including other Thai samples) at this threshold, indicating that this
319 threshold was too liberal at splitting species.
320 mPTP.—the mPTP analysis did not support the splitting of balingiana and
321 malaccae (PP=0.16); bearni Borneo and bearni PM (PP=0.13); or tener Borneo and
322 tener PM east (PP=0.67); but highly supported the split between tener PM west and
323 tener Borneo + tener PM east (PP=1.0).
324 BPP and gdi.—all candidate populations were delimited as distinct species
325 with high support (posterior probability=1.0) in the BPP A10 and A11 analyses (Table
326 2). The heuristic gdi analysis provided moderately high but uncertain support for the
327 split between balingiana and malaccae; tener Borneo and tener PM east; and bearni
328 Borneo and bearni PM. The split between tener PM west and tener Borneo + tener
329 PM east was highly supported (Fig. 4; Table 2).
330
331 DISCUSSION
332 Phylogenetic relationships
11
333 Phylogenetic relationships in Luciolinae were exclusively based on ~400

334 morphological characters (Ballantyne & Lambkin, 2013; Ballantyne et al., 2015) until
335 Jusoh et al. (2018) made the first attempt to combine and jointly analyze the
336 morphological data with a small subset of molecular data from 25 taxa (total = 158
337 taxa) using Bayesian Inference (BI) and Maximum Parsimony (MP). Although taxon
338 sampling was limited, the molecular data recovered a similar topology with the larger
339 morphological matrix, but the combined morphological and molecular dataset
340 analysis produced different relationships. Within the Pteroptyx clade, the MP analysis
341 (Figure 2, part 2 in Jusoh et al. 2018) revealed that P. balingiana formed a clade with
342 Pteroptyx macdermotti McLean and P. gelasina and this clade was reciprocally
343 monophyletic with P. malaccae (the BI analysis was unresolved). This does not
344 conflict with the topology from this study as no molecular samples of P. macdermotti
345 and P. gelasina were available for analysis. However, the phylogenetic position of P.
346 valida and P. asymmetria were not concordant with results from our analyses. Our
347 analysis recovered P. asymmetria as the basal lineage with regard to P. valida, P.
348 malaccae, P. bearni, and P. tener, but Jusoh et al (2018) recovered it within the P.
349 bearni + P. tener subclade, which was in turn sister to the P. valida clade. Based on
350 morphology, P. asymmetria is more affine to P. tener and P. bearni. Our study
351 recovered P. valida as the sister lineage to the P. malaccae + P. balingiana subclade
352 albeit with moderate support. Unfortunately at this point, we are still unable to resolve
353 the species relationships of Malaysia Pteroptyx with any amount of certainty and
354 more data in the form of genetic loci and taxon representation will be needed to
355 provide more robust inferences.
356
357 Species boundaries and systematics
358 Using a suite of species delimitation analyses that are based on different models and
359 assumptions, we showed that Pteroptyx from Southeast Asia runs the gamut of the
360 speciation continuum and exhibits strong phylogeographic patterns that are in some
361 cases incongruent with morphological differentiation. Populations of P. tener from
362 Borneo + PM east showed high genetic divergence and was highly supported as
363 distinct from the PM west population across all analyses. Furthermore, sampling for
364 each of these populations were robust and represented by multiple individuals. As
12
365 such, the results can be considered robust and the specific distinction between these
366 lineages can be posited with confidence. However, there are no reliable diagnostic
367 characters that distinguish these lineages (see Fig. 5A; Jusoh et al., 2018), indicating
368 that they constitute cryptic species (Bickford et al., 2007).
369 Pteroptyx balingiana was described as a separate species from P. malaccae

370 on the basis of morphological and molecular differentiation (Jusoh et al., 2018).
371 Morphologically, P. balingiana can be distinguished from P. malaccae by structural
372 differences in the light emitting organ (Fig. 5B,C). More in-depth analyses performed
373 here demonstrated that genetic divergence between the two taxa was moderate and
374 their status as distinct species was not unequivocally supported by species
375 delimitation analyses. Nevertheless, the data also did not strongly support the
376 lumping of these taxa. In light of the following lines of evidence: i) presence of
377 morphological differences; ii) allopatry; iii) genetic differentiation (albeit modest);
378 these taxa clearly represent independently evolving metapopulation lineages that can
379 be considered as distinct species under the General Lineage Concept of Species (de
380 Queiroz, 2007). As such, we retain the current taxonomic status of P. balingiana and
381 P. malaccae as distinct species until additional data suggest otherwise.
382 Similarly, species delimitation results for populations of P. bearni from Borneo/
383 PM and populations of P. tener from Borneo/PM east were also largely uncertain.
384 According to the morphological examination of Jusoh et al. (2018) which consist of
385 freshly preserved specimens, individuals of P. bearni from PM and Borneo can be
386 distinguished by the colouration on the pronotum and head (Fig. 5D). However, these
387 differences were not observable in preserved specimens. No morphological
388 differences were detected between populations of P. tener from Borneo and eastern
389 PM (Fig. 5A), but this could be due to the limited number of Bornean samples. Based
390 on the relatively shallow bipartitions of these lineages and the absence or subtle
391 differences in morphology, we interpret them as being in the early stages of
392 speciation where insufficient time has elapsed for considerable genetic differences to
393 accumulate. Demonstrably, the distribution of genetic divergences show slight but
394 clear shifts away from population-level variation, but have yet to attain the levels of
395 divergence observed between species (Fig. 2). We therefore consider these lineages
396 as incipient species that have begun to diverge, but remain in the gray zone of the
397 speciation continuum (Feder, Egan, & Nosil, 2012; Nosil & Feder, 2012; Roux et al.,
13
398 2016). Because evidence was not compelling, we refrain from splitting these
399 populations into separate species pending further investigation.
400
401 Caveats, limitations, and future directions
402 Both mPTP and bGMYC methods seek to identify the transition point between the
403 speciation and coalescent process, hence adequate taxon representation is needed
404 to provide robust estimates. Consequently, populations represented by singletons or
405 doubletons can affect the accuracy of the analysis (Kapli et al., 2017; Reid &
406 Carstens, 2012). While BPP has been shown to be less sensitive to taxon rarity or
407 low numbers of loci (Yang & Rannala, 2017), this analysis is also known to oversplit
408 species due to its inability to differentiate between population structure and species
409 divergence under a protracted speciation model (Chan & Grismer, 2019; Leaché et
410 al., 2018; Sukumaran & Knowles, 2017). This could explain the unrealistically high
411 support for splitting every one of the focal populations, which is in conflict with results
412 from the other species delimitation analyses. Therefore, the BPP analyses and
413 results pertaining to populations with singletons or doubletons should be treated with
414 caution.
415 The limited number of genes that was available for analysis could also
416 potentially affect the results. Unfortunately, there is a severe lack of nuclear gene
417 representation in public databases. Furthermore, additional studies are required to
418 identify informative genes that can be useful for phylogenetics and species
419 delimitation in fireflies, as this has yet to be determined. There is also a dearth of
420 taxonomic representation in molecular samples where only eight out of 18 species of
421 Pteroptyx have ever been sequenced, rendering the phylogenetic relationships of this
422 genus uncertain. Morphologically, Pteroptyx is a genus with three subdivisions:
423 Group I (deflexed elytra + meta femoral comb (MFC) + bipartite light organ (BLO))
424 (see Fig. 5); Group II (non-deflexed elytra + MFC + entire light organ (ELO)); and
425 Group III (non-deflexed elytra + no MFC + BLO; see Jusoh et al. 2018). This study
426 only dealt with taxa from Group I or commonly known as the bent-winged fireflies,
427 with poor representative from the other groups (Groups II = P. galbina, III = P.
428 testacea). Future phylogenetic studies should ideally include key taxa that are closely
429 related to Pteroptyx such as Medeopteryx Ballantyne and Pyrophanes Olivier.
14
430 Another species that is morphologically similar to P. malaccae and P. balingiana is P.

431 gelasina, which is only known from Sabah, Borneo. Unfortunately, no molecular data
432 is available for this species. Therefore, apart from increasing the genetic sampling of
433 P. balingiana, acquiring molecular data for P. gelasina is also crucial to determine the
434 species boundaries and phylogeographic structure of this species complex. While
435 missing taxa could affect the topology of the overall phylogeny, we do not anticipate
436 that it will significantly disrupt our species delimitation results as those were
437 performed at the population level (i.e. the populations within the P. malaccae, P.
438 bearni, and P. tener groups), which are not affected by uncertainty in relationships at
439 the species level.
440
441 Biogeography
442 Due to the limited number of genes and the absence of calibration points, we were
443 unable to anchor the phylogeny onto a temporal framework of absolute time and this
444 prevented us from providing an estimate on diversification times. However, based on
445 the relatively shallow divergences among populations from PM and Borneo, we
446 hypothesize that cyclical glaciations during the Pleistocene (Hall, 2013) could have
447 facilitated the vicariance of these populations, resulting in the fragmented present-
448 day distribution on their respective land masses (Erik, 2003; Lim et al., 2017; Slik et
449 al., 2011; Wurster et al., 2010). Within Peninsular Malaysia, the east-west pattern of
450 lineage separation has been demonstrated in numerous co-distributed taxa (Chan,
451 Abraham, Grismer, & Grismer, 2018; Chan et al., 2017, 2014; Grismer et al., 2013,
452 2015). The isolation of those taxa were hypothesized to be caused by the relatively
453 central and contiguous Titiwangsa mountain range, which acts as an insurmountable
454 physical barrier that separates habitats along the eastern and western regions (Chan
455 et al., 2017; Chan & Brown, 2019). We consider this proposed model to be
456 incompatible with Pteroptyx, which occurs in mangroves and do not disperse across
457 inland habitats to begin with. In this scenario, inland habitats represent ecological
458 barriers and dispersal routes are consequently restricted to the coastline. However,
459 the distribution of ancient mangroves are likely to be different compared to the
460 present day and a recent study suggested that the dispersal of certain populations of
15
461 P. tener in Peninsular Malaysia could have occurred along ancient river networks
462 during the Pleistocene (Cheng, Munian, Sek-Aun, Faidi, & Ishak, 2019).
463
464 CONCLUSIONS
465 Despite the charismatic attraction that fireflies hold in cultures all around the world,
466 this study revealed that fundamental knowledge on Southeast Asian firefly
467 biodiversity and evolutionary history is still lacking. More disconcertingly, much of
468 their habitat, specifically forests and mangroves are suffering severe degradation and
469 loss from pollution and land conversion (Jusoh & Hashim, 2012; Prasertkul, 2018;
470 Wong & Yeap, 2012), setting the stage for an arms race to elucidate the true extent
471 of firefly diversity before it is lost. Our data showed that Pteroptyx fireflies in Malaysia
472 contain hitherto undiscovered cryptic and incipient species that are
473 phylogeographically structured, thereby demonstrating the dynamic and complicated
474 interplay between genes and the environment that drives, maintains, and distributes
475 firefly biodiversity. Overall, this study contributes to the growing body of genetic work
476 on fireflies and underscores the urgent need to increase the breadth and depth of
477 geographic and genetic sampling to better understand the evolutionary history of
478 fireflies.
479
480 ACKNOWLEDGEMENTS
481 WFA thanks the Evolutionary Biology Lab at the National University of Singapore for
482 support on molecular work.
483
484
485
486
487
488
489
16
490 REFERENCES
491 Ballantyne, L. (1987). Further Revisional Studies on the Firefly Genus Pteroptyx
492 Olivier ( Coleoptera: Lampyridae: Luciolinae: Luciolini ). Transactions of the
493 American Entomological Society, 113(2), 117–170. Retrieved from
494 http://www.jstor.org/stable/10.2307/25078409
495 Ballantyne, L. (2001). The bent winged fireflies of Cambodia, Indonesia, Malaysia,
496 Philippines and Thailand (Coleoptera: Lampyridae: Luciolinae: Luciolini).
497 Pteroptyx spp. of the Polunin collection. Serangga, 6(1), 51–95.
498 Ballantyne, L. A., & Lambkin, C. L. (2013). Systematics and Phylogenetics of Indo-
499 Pacific Luciolinae Fireflies (Coleoptera: Lampyridae) and the Description of new
500 Genera. In Zootaxa (Vol. 3653). https://doi.org/10.11646/zootaxa.3653.1.1
501 Ballantyne, L. A., & McLean, M. R. (1970). Revisional studies on the firefly genus
502 Pteroptyx Olivier (Coleoptera: Lampyridae: Luciolinae: Luciolini). Transactions of
503 the American Entomological Society, 96(August 18,1970), 223–305. Retrieved
504 from http://www.jstor.org/stable/25077994
505 Ballantyne, L., Fu, X. H., Shih, C. H., Cheng, C. Y., & Yiu, V. (2011). Pteroptyx maipo
506 ballantyne, a new species of bent-winged firefly (coleoptera: Lampyridae) from
507 Hong Kong, and its relevance to firefly biology and conservation. Zootaxa,
508 34(2931), 8–34.
509 Ballantyne, L., Lambkin, C. L., Boontop, Y., & Jusoh, W. F. A. (2015). Revisional
510 studies on the luciolinae fireflies of Asia (Coleoptera: Lampyridae): 1. The genus
511 pyrophanes olivier with two new species. 2. Four new species of pteroptyx olivier
512 and 3. A new genus inflata boontop, with redescription of Luciola indica (Motsc.
513 Zootaxa, 3959(1), 1–84. https://doi.org/10.11646/zootaxa.3959.1.1
514 Bickford, D., Lohman, D. J., Sodhi, N. S., Ng, P. K. L., Meier, R., Winker, K., … Das, I.
515 (2007). Cryptic species as a window on diversity and conservation. Trends in
516 Ecology and Evolution, Vol. 22, pp. 148–155.
517 https://doi.org/10.1016/j.tree.2006.11.004
518 Bouckaert, R., Heled, J., Kühnert, D., Vaughan, T., Wu, C. H., Xie, D., … Drummond,
519 A. J. (2014). BEAST 2: A Software Platform for Bayesian Evolutionary Analysis.
520 PLoS Computational Biology, 10(4), 1–6.
17
521 https://doi.org/10.1371/journal.pcbi.1003537
522 Bouckaert, R. R., & Drummond, A. J. (2017). bModelTest: Bayesian phylogenetic site
523 model averaging and model comparison. BMC Evolutionary Biology, 17(1), 42.
524 https://doi.org/10.1101/020792
525 Buck, J. (2004). Synchronous Rhythmic Flashing of Fireflies. II. The Quarterly
526 Review of Biology, 63(3), 265–289. https://doi.org/10.1086/415929
527 Cao, X., Liu, J., Chen, J., Zheng, G., Kuntner, M., & Agnarsson, I. (2016). Rapid
528 dissemination of taxonomic discoveries based on DNA barcoding and
529 morphology. Scientific Reports, 6(1), 1–13. https://doi.org/10.1038/srep37066
530 Case, J. F. (2007). Courting Behavior in a Synchronously Flashing, Aggregative

531 Firefly, Pteroptyx Tener. The Biological Bulletin, 159(3), 613–625.
532 https://doi.org/10.2307/1540827
533 Chan, K. O., Abraham, R. K., Grismer, J. L., & Grismer, L. L. (2018). Elevational size
534 variation and two new species of torrent frogs from Peninsular Malaysia (Anura:
535 Ranidae: Amolops Cope). Zootaxa, 4434(2), 250–264.
536 Chan, K. O., Alexander, A. M., Grismer, L. L., Su, Y.-. C., Grismer, J. L., & Quah, E.
537 S. H. (2017). Species delimitation with gene flow: a methodological comparison
538 and population genomics approach to elucidate cryptic species boundaries in
539 Malaysian torrent frogs. Mol Ecol, 26.
540 Chan, K. O., & Brown, R. M. (2019). Linking patterns of genetic variation to
541 processes of diversification in Malaysian torrent frogs (Anura: Ranidae: BioArxiv,
542 628891. https://doi.org/https://doi.org/10.1101/628891
543 Chan, K. O., & Grismer, L. L. (2019). To split or not to split? Multilocus phylogeny
544 and molecular species delimitation of southeast Asian toads (family: Bufonidae).
545 BMC Evolutionary Biology, 19(1), 95. https://doi.org/10.1186/s12862-019-1422-3
546 Chan, K. O., Grismer, L. L., & Brown, R. M. (2018). Comprehensive multi-locus
547 phylogeny of Old World tree frogs (Anura: Rhacophoridae) reveals taxonomic
548 uncertainties and potential cases of over- and underestimation of species
549 diversity. Molecular Phylogenetics and Evolution, 127, 1010–1019.
550 https://doi.org/10.1016/j.ympev.2018.07.005
18
551 Chan, K. O., Wood, P. L. J., Anuar, S., Muin, M. A., Quah, E. S. H., Sumarli, A. X. Y.,
552 & Grismer, L. L. (2014). A new species of upland Stream Toad of the genus
553 Ansonia Stoliczka, 1870 (Anura: Bufonidae) from northeastern Peninsular
554 Malaysia. Zootaxa, 3764(4), 427–440.
555 Chan, S. H., Ballantyne, L. A., & Goh, C. A. L. (2012). Survey of species diversity ,
556 distribution and abundance of fireflies in Singapore. Lampyrid, 113–122.
557 Cheng, S., Munian, K., Sek-Aun, T., Faidi, M. A., & Ishak, S.-F. (2019). Mitochondrial
558 DNA diversity and gene flow in Southeast Asian populations of the
559 synchronously flashing firefly, Pteroptyx tener Olivier (Coleoptera: Lampyridae).
560 Oriental Insects, 00(00), 1–22. https://doi.org/10.1080/00305316.2019.1600594
561 Copeland, J., & Moiseff, A. (2006). Flash Precision at the Start of Synchrony in
562 Photuris frontalis. Integrative and Comparative Biology, 44(3), 259–263.
563 https://doi.org/10.1093/icb/44.3.259
564 de Queiroz, K. (2007). Species concepts and species delimitation. Systematic

565 Biology, 56(6), 879–886. https://doi.org/10.1080/10635150701701083
566 Erik, M. (2003). Mammals of south-east Asian islands and their Late Pleistocene
567 environments. Journal of Biogeography, 30(8), 1245–1257.
568 Ermentrout, B. (1991). An adaptive model for synchrony in the firefly Pteroptyx
569 malaccae. Journal of Mathematical Biology, 29(6), 571–585.
570 https://doi.org/10.1007/BF00164052
571 Faust, L. F. (2004). Fireflies as a catalyst for science education. Integrative and
572 Comparative. Biology, 44(3), 264–265. Retrieved from
573 http://www.jstor.org/stable/3884718
574 Feder, J. L., Egan, S. P., & Nosil, P. (2012). The genomics of speciation-with-gene-
575 flow. Trends in Genetics, 28(7), 342–350.
576 https://doi.org/10.1016/j.tig.2012.03.009
577 Flouri, T., Jiao, X., Rannala, B., & Yang, Z. (2018). Species tree inference with BPP
578 using genomic sequences and the multispecies coalescent. Molecular Biology
579 and Evolution, 35(10), 2585–2593. https://doi.org/10.1093/molbev/msy147
580 Folmer, O., Black, M., Hoeh, W., Lutz, R., & Vrijenhoek, R. (1994). DNA primers for
19
581 amplification of mitochondrial cytochrome c oxidase subunit I from diverse

582 metazoan invertebrates. Molecular Marine Biology and Biotechnology, 3(5),
583 294–299.
584 Foo, K., & Dawood, M. M. (2017). Diversity of Pteroptyx Fireflies (Coleoptera:
585 Lampyridae) and Their Display Trees at Klias Peninsula, Sabah, Malaysia.
586 Journal of Tropical Biology and Conservation, 14, 95–103. Retrieved from
587 https://www.ums.edu.my/ibtpv2/files/07.pdf
588 Fouquet, A., Gilles, A., Vences, M., Marty, C., Blanc, M., & Gemmell, N. J. (2007).
589 Underestimation of species richness in neotropical frogs revealed by mtDNA
590 analyses. PLoS ONE, 2(10). https://doi.org/10.1371/journal.pone.0001109
591 Fraga, H. (2008). Firefly luminescence: A historical perspective and recent

592 developments. Photochemical and Photobiological Sciences, 7(2), 146–158.
593 https://doi.org/10.1039/b719181b
594 Gould Stephen J., & Subramani Suresh. (1988). Firefly Luciferase as a Tool in
595 Molecular and Cell Biology. Analytical Biochemistry, 175, 5–13.
596 Grismer, L. L., Wood, P. L., Anuar, S., Muin, M. A., Quah, E. S. H., McGuire, J. A., …
597 Pham, H. T. (2013). Integrative taxonomy uncovers high levels of cryptic species
598 diversity in Hemiphyllodactylus Bleeker, 1860 (Squamata: Gekkonidae) and the
599 description of a new species from Peninsular Malaysia. Zoological Journal of the
600 Linnean Society, 169(4), 849–880. https://doi.org/10.1111/zoj.12064
601 Grismer, L. L., Wood, P. L., Anuar, S., Quah, E. S. H., Muin, M. A., Chan, K. O., …
602 Loredo, A. I. (2015). Repeated evolution of sympatric, palaeoendemic species in
603 closely related, co-distributed lineages of Hemiphyllodactylus Bleeker, 1860
604 (Squamata: Gekkonidae) across a sky-island archipelago in Peninsular Malaysia.
605 Zoological Journal of the Linnean Society, 174(4), 859–876.
606 https://doi.org/10.1111/zoj.12254
607 Hall, R. (2013). The palaeogeography of Sundaland and Wallacea since the Late
608 Jurassic. Journal of Limnology, 72(S2), 1–17.
609 https://doi.org/10.4081/jlimnol.2013.s2.e1
610 Hebert, P. D. N., Cywinska, A., Ball, S. L., & DeWaard, J. R. (2003). Biological
611 identifications through DNA barcodes. Proceedings of the Royal Society B:
20
612 Biological Sciences, 270(1512), 313–321.

613 https://doi.org/10.1098/rspb.2002.2218
614 Hebert, P. D. N., & Gregory, T. R. (2005). The Promise of DNA Barcoding for
615 Taxonomy. Systematic Biology, Vol. 54, pp. 852–859.
616 https://doi.org/10.1080/10635150500354886
617 Hoang, D. T., Chernomor, O., Von Haeseler, A., Minh, B. Q., & Vinh, L. S. (2018).
618 UFBoot2: Improving the ultrafast bootstrap approximation. Molecular Biology
619 and Evolution, 35(2), 518–522. https://doi.org/10.1093/molbev/msx281
620 Hubert, N., & Hanner, R. (2016). DNA Barcoding, species delineation and taxonomy:
621 a historical perspective. DNA Barcodes, 3(1), 44–58.
622 https://doi.org/10.1515/dna-2015-0006
623 Jackson, N. D., Carstens, B. C., Morales, A. E., & O’Meara, B. C. (2017a). Species
624 delimitation with gene flow. Syst Biol, 66.
625 Jackson, N. D., Carstens, B. C., Morales, A. E., & O’Meara, B. C. (2017b). Species
626 delimitation with gene flow. Systematic Biology, 66(5), 799–812.
627 https://doi.org/10.1093/sysbio/syw117
628 Jusoh, W. F. A., Ballantyne, L., Lambkin, C. L., Hashim, N. R., & Wahlberg, N. (2018).
629 The firefly genus Pteroptyx Olivier revisited (Coleoptera: Lampyridae: Luciolinae).
630 Zootaxa, 4456(1), 1–71. https://doi.org/10.11646/zootaxa.4456.1.1
631 Jusoh, W. F. A., Hashim, N. R., Sääksjärvi, I. E., Adam, N. A., & Wahlberg, N. (2014).
632 Species Delineation of Malaysian Mangrove Fireflies (Coleoptera: Lampyridae)
633 using DNA Barcodes. The Coleopterists Bulletin, 68(4), 703–711.
634 https://doi.org/10.1649/0010-065x-68.4.703
635 Jusoh, W. F. A. W., & Hashim, N. R. (2012). The effect of habitat modification on
636 firefly populations at the Rembau-Linggi estuary, Peninsular Malaysia. Lampyrid,
637 2, 149–155.
638 Jusoh, W. F. A. W., Hashim, N. R., & Ibrahim, Z. Z. (2010). Firefly distribution and
639 abundance on mangrove vegetation assemblages in Sepetang estuary,
640 Peninsular Malaysia. Wetlands Ecology and Management, 18(3), 367–373.
641 https://doi.org/10.1007/s11273-009-9172-4
21
642 Kalyaanamoorthy, S., Minh, B. Q., Wong, T. K. F., Von Haeseler, A., & Jermiin, L. S.
643 (2017). ModelFinder: Fast model selection for accurate phylogenetic estimates.
644 Nature Methods, 14(6), 587–589. https://doi.org/10.1038/nmeth.4285
645 Kapli, P., Lutteropp, S., Zhang, J., Kobert, K., Pavlidis, P., Stamatakis, A., & Flouri, T.
646 (2017). Multi-rate Poisson tree processes for single-locus species delimitation
647 under maximum likelihood and Markov chain Monte Carlo. Bioinformatics,
648 33(11), 1630–1638. https://doi.org/10.1093/bioinformatics/btx025
649 Kumar, S., Stecher, G., Li, M., Knyaz, C., & Tamura, K. (2018). MEGA X: Molecular
650 evolutionary genetics analysis across computing platforms. Molecular Biology
651 and Evolution, 35(6), 1547–1549. https://doi.org/10.1093/molbev/msy096
652 Leaché, A. D., Zhu, T., Rannala, B., & Yang, Z. (2018). The Spectre of Too Many
653 Species. Systematic Biology, 0(0), 1–14. https://doi.org/10.1093/sysbio/syy051
654 Lim, H. C., Gawin, D. F., Shakya, S. B., Harvey, M. G., Rahman, M. A., & Sheldon, F.
655 H. (2017). Sundaland’s east-west rain forest population structure: Variable
656 manifestations in four polytypic bird species examined using RAD-Seq and
657 plumage analyses. Journal of Biogeography. https://doi.org/10.1111/jbi.13031
658 Lloyd, J. E. (2002). Lampyridae. In R. A. Arnett, M. C. Thomas, P. E. Skelley, & J. H.

659 Frank (Eds.), American Beetles Vol. 2 (pp. 187–196). CRC Press, Boca Raton.
660 Lloyd, J. E. (2008). Fireflies (Coleoptera: Lampyridae). In Encyclopedia of

661 Entomology (pp. 1429–1452). https://doi.org/10.1007/978-1-4020-6359-6_3811
662 Lloyd, J. E., Wing, S. R., & Hongtrakul, T. (2006). Flash Behavior and Ecology of
663 Thai Luciola Fireflies (Coleoptera: Lampyridae). The Florida Entomologist, 72(1),
664 80. https://doi.org/10.2307/3494969
665 McDermott, F. A. (1966). Lampyridae. In W. O. Steel (Ed.), Coleopterorum catalogus

666 supplementa, pars 9. (pp. 1–149). Gravenhage, Netherlands: Uitgeverij Dr. W.
667 Junk.
668 Miller, M. A., Pfeiffer, W., & Schwartz, T. (2010). Creating the CIPRES Science
669 Gateway for inference of large phylogenetic trees. 2010 Gateway Computing
670 Environments Workshop, GCE 2010, 1–8.
671 https://doi.org/10.1109/GCE.2010.5676129
22
672 Nguyen, L. T., Schmidt, H. A., Von Haeseler, A., & Minh, B. Q. (2015). IQ-TREE: A
673 fast and effective stochastic algorithm for estimating maximum-likelihood
674 phylogenies. Molecular Biology and Evolution, 32(1), 268–274.
675 https://doi.org/10.1093/molbev/msu300
676 Nosil, P., & Feder, J. L. (2012). Genomic divergence during speciation: Causes and
677 consequences. Philosophical Transactions of the Royal Society B: Biological
678 Sciences, 367(1587), 332–342. https://doi.org/10.1098/rstb.2011.0263
679 Padial, J. M., & De La Riva, I. (2007). Integrative taxonomists should use and
680 produce DNA barcodes. Zootaxa. https://doi.org/10.11646/zootaxa.1586.1.7
681 Pinho, C., & Hey, J. (2010). Divergence with Gene Flow: Models and Data. Annual
682 Review of Ecology, Evolution, and Systematics, 41(1), 215–230.
683 https://doi.org/10.1146/annurev-ecolsys-102209-144644
684 Prasertkul, T. (2018). Characteristics of Pteroptyx Firefly Congregations in a Human

685 Dominated Habitat. Journal of Insect Behavior, 31(4), 436–457.
686 https://doi.org/10.1007/s10905-018-9687-8
687 Puillandre, N., Lambert, A., Brouillet, S., & Achaz, G. (2012). ABGD, Automatic
688 Barcode Gap Discovery for primary species delimitation. Molecular Ecology,
689 21(8), 1864–1877. https://doi.org/10.1111/j.1365-294X.2011.05239.x
690 Rambaut, A., Drummond, A. J., Xie, D., Baele, G., & Suchard, M. A. (2018). Posterior
691 summarization in Bayesian phylogenetics using Tracer 1.7. Systematic Biology,
692 67(5), 901–904. https://doi.org/10.1093/sysbio/syy032
693 Ratnasingham, S., & Hebert, P. D. N. (2013). A DNA-Based Registry for All Animal
694 Species: The Barcode Index Number (BIN) System. PLoS ONE, 8(7).
695 https://doi.org/10.1371/journal.pone.0066213
696 Reid, N. M., & Carstens, B. C. (2012). Phylogenetic estimation error can decrease
697 the accuracy of species delimitation: A Bayesian implementation of the general
698 mixed Yule-coalescent model. BMC Evolutionary Biology, 12(1), 196.
699 https://doi.org/10.1186/1471-2148-12-196
700 Roux, C., Fraïsse, C., Romiguier, J., Anciaux, Y., Galtier, N., & Bierne, N. (2016).
701 Shedding Light on the Grey Zone of Speciation along a Continuum of Genomic
23
702 Divergence. PLoS Biology, 14(12), 1–22.

703 https://doi.org/10.1371/journal.pbio.2000234
704 Sartsanga, C., Swatdipong, A., & Sriboonlert, A. (2018). Distribution of the Firefly
705 Genus Pteroptyx Olivier and a New Record of Pteroptyx asymmetria Ballantyne
706 (Coleoptera: Lampyridae: Luciolinae) in Thailand . The Coleopterists Bulletin,
707 72(1), 171–183. https://doi.org/10.1649/0010-065x-72.1.171
708 Schena, A., Griss, R., & Johnsson, K. (2015). Modulating protein activity using
709 tethered ligands with mutually exclusive binding sites. Nature Communications,
710 6. https://doi.org/10.1038/ncomms8830
711 Sheth, B. P., & Thaker, V. S. (2017). DNA barcoding and traditional taxonomy: an
712 integrated approach for biodiversity conservation. Genome, 60(7), 618–628.
713 https://doi.org/10.1139/gen-2015-0167
714 Sites, J. W., Marshall, J. C., Jr, J. W. S., & Marshall, J. C. (2003). Delimiting species:
715 a renaissance issue in systematic biology. Trends in Ecology and Evolution,
716 18(9), 462–470. https://doi.org/10.1016/S0169-5347(03)00184-8
717 Slik, J. W. F., Aiba, S.-I., Bastian, M., Brearley, F. Q., Cannon, C. H., Eichhorn, K. A.
718 O., … Zweifel, N. (2011). Soils on exposed Sunda Shelf shaped biogeographic
719 patterns in the equatorial forests of Southeast Asia. Proceedings of the National
720 Academy of Sciences, 108(30), 12343–12347.
721 https://doi.org/10.1073/pnas.1103353108
722 Sriboonlert, A., Swatdipong, A., Wonnapinij, P., E-Kobon, T., & Thancharoen, A.
723 (2015). New Record of Pteroptyx tener Olivier (Coleoptera: Lampyridae:
724 Luciolinae) in Thailand . The Coleopterists Bulletin, 69(2), 332–336.
725 https://doi.org/10.1649/0010-065x-69.2.332
726 Sukumaran, J., & Knowles, L. L. (2017). Multispecies coalescent delimits structure,
727 not species. Proceedings of the National Academy of Sciences, 114(7), 1607–
728 1612. https://doi.org/10.1073/pnas.1607921114
729 Tang, C. Q., Humphreys, A. M., Fontaneto, D., & Barraclough, T. G. (2014). Effects
730 of phylogenetic reconstruction method on the robustness of species delimitation
731 using single-locus data. Methods in Ecology and Evolution, 5(10), 1086–1094.
732 https://doi.org/10.1111/2041-210X.12246
24
733 Vences, M., Thomas, M., Bonett, R. M., & Vieites, D. R. (2005). Deciphering
734 amphibian diversity through DNA barcoding: Chances and challenges.
735 Philosophical Transactions of the Royal Society B: Biological Sciences,
736 360(1462), 1859–1868. https://doi.org/10.1098/rstb.2005.1717
737 Vences, M., Thomas, M., Meijden, A. Van Der, Chiari, Y., & Vieites, D. R. (2005). of
738 Amphibians. Frontiers in Zoology, 12, 1–12. https://doi.org/10.1186/1742-9994-
739 2-5
740 Vieites, D. R., Wollenberg, K. C., Andreone, F., Kohler, J., Glaw, F., & Vences, M.
741 (2009). Vast underestimation of Madagascar’s biodiversity evidenced by an
742 integrative amphibian inventory. Proceedings of the National Academy of
743 Sciences, 106(20), 8267–8272. https://doi.org/10.1073/pnas.0810821106
744 Villalba, S., Lobo, J. M., Martin-Piera, F., & Zardoya, R. (2002). Phylogenetic
745 relationships of Iberian Dung Beetles (Coleoptera: Scarabaeinae): insights on
746 the evlution of nesting behavior. Journal of Molecular Evolution, 55, 116–126.
747 Wiens, J. J. (2007). Species Delimitation: New Approaches for Discovering Diversity.
748 Systematic Biology, 56(6), 875–878.
749 https://doi.org/10.1080/10635150701748506
750 Wong, C.-H., & Yeap, C. A. (2012). Conservation of Congregating Firefly Zones
751 (CFZs) in Peninsular Malaysia. Lampyrid, 2, 174–187.
752 Wurster, C. M., Bird, M. I., Bull, I. D., Creed, F., Bryant, C., Dungait, J. A. J., & Paz, V.
753 (2010). Forest contraction in north equatorial Southeast Asia during the Last
754 Glacial Period. Proceedings of the National Academy of Sciences, 107(35),
755 15508–15511. https://doi.org/10.1073/pnas.1005507107
756 Yang, Z., & Rannala, B. (2010). Bayesian species delimitation using multilocus
757 sequence data. Proceedings of the National Academy of Sciences, 107(20),
758 9264–9269. https://doi.org/10.1073/pnas.0913022107
759 Yang, Z., & Rannala, B. (2017). Bayesian species identification under the
760 multispecies coalescent provides significant improvements to DNA barcoding
761 analyses. Molecular Ecology, 26(11), 3028–3036.
762 https://doi.org/10.1111/mec.14093
25
763 Zhang, J., Kapli, P., Pavlidis, P., & Stamatakis, A. (2013). A general species
764 delimitation method with applications to phylogenetic placements. Bioinformatics,
765 29(22), 2869–2876. https://doi.org/10.1093/bioinformatics/btt499
766 Ziheng, Y. (2015). The BPP program for species tree estimation and species
767 delimitation. Current Zoology, 61(5), 854–865.
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
26
788 FIGURE LEGENDS
789 Fig. 1. Geographic distribution of Pteroptyx samples used in this study and the
790 BEAST phylogeny inferred from a concatenated sequence matrix consisting of 2,119
791 bps of the CO1 mitochondrial and CAD nuclear genes. Circles at nodes denote
792 Bayesian posterior probabilities and ML Ultrafast Bootstrap support values. Nodes
793 without circles have low support (PP<0.9; UFB<95). Species-level clades are color-
794 coded to match locality points on the distribution map. Inset figures show dorsal and
795 ventral views of male representative species from the ingroup.
796
797 Fig. 2. Distribution of intra and interspecific genetic distances based on the CO1
798 mitochondrial gene. Candidate species represent reciprocally monophyletic and
799 geographically circumscribed populations that were subjected to downstream species
800 delimitation analyses. Dotted red lines represent the 3%, 4%, and 5% genetic
801 distance thresholds.
802
803 Fig. 3. Results of the initial and recursive runs of ABGD and the number of delimited
804 species at the associated maximal intraspecific distance (P).
805
806 Fig. 4. Density plots of gdi values with vertical dotted lines representing the 0.2 and
807 0.7 thresholds. Populations are considered distinct species if gdi >0.7, while gdi <0.2
808 indicate that populations belong to the same species. Values of 0.2> gdi < 0.7
809 indicate ambiguous species status.
810
811 Fig. 5A. Ventral view of Pteroptyx tener from PM east, PM west, and Borneo; B and
812 C. dorsal and ventral view of P. balingiana and P. malaccae (note the difference in
813 the shape of the light organ); D. P. bearni from PM east and Borneo can be
814 differentiated by the colouration of the head but both populations have similar light
815 organ structures.
816
27
817 SUPPORTING INFORMATION
818 Fig. S1. Maximum likelihood phylogeny estimated using a partitioned analysis on a
819 concatenated sequence matrix consisting of 2,119 bps of the CO1 mitochondrial and
820 CAD nuclear genes. Node values denote ultrafast bootstrap support.
821
822 Fig. S2. Bayesian phylogeney inferred from a concatenated sequence matrix
823 consisting of 2,119 bps of the CO1 mitochondrial and CAD nuclear genes. Node
824 values denote posterior probabilities.
825
826 Table S1. List of genetic samples used in this study and their corresponding
827 GenBank accession numbers.
828
829 Supporting information 1. Aligned sequences for the CO1 mitochondrial gene.
830
831 Supporting information 2. Aligned sequences for the CAD nuclear gene.
832
833
834
835
836
837
838
839
840
841
842
28
843 TABLES
844 Table 1. Partitions, number of species (initial run followed by recursive run in
845 parenthesis), and corresponding prior maximal distance from the ABGD analysis
846 using the Kimura (K80) distance model and a relative gap with of X=1.5.
Number of species Prior max. distance
Partition 1 10(26) 0.001
Partition 2 10(14) 0.001668
Partition 3 10(13) 0.002783
Partition 4 10(11) 0.004642
Partition 5 10(10) 0.007743
Partition 6 9(10) 0.012915
Partition 7 1(1) 0.021544
847
848
849 Table 2. Summary of the species delimitation results on the candidate species.
balingiana vs. bearni Borneo tener Borneo tener Borneo+PM east
malaccae vs. bearni PM vs. tener PM east vs. tener PM west
P- distance (%) 1.48–2.92 0.84–5.00 1.39–2.29 3.15–5.49
ABGD Lump/split Lump/split Lump/split Split
mPTP Low (0.16) Low (0.13) Low (0.67) High (1.0)
bGMYC Low (0.29) Low (0.23) Low (0.23) Moderately High (0.08)
BPP (A10) High (1.0) High (1.0) High (1.0) High (1.0)
BPP (A11) High (1.0) High (1.0) High (1.0) High (1.0)
gdi High Uncertain Uncertain High
850
29
851 Combined results from partitions 4–6 that delimited 9–10 species. See Results for
852 more details.
853 Support values denote posterior probabilities that lineages are separate species.
854 PP > 0.95=High; 0.9<PP<0.95 = Moderate; PP < 0.9 = Low.
855 Support values denote posterior probabilities that lineages are conspecific. Hence,
856 lineages are highly supported as separate species if PP <0.05, while 0.1 > PP > 0.05
857 = moderate support; PP > 1.0 = Low support.
858 Average gdi > 0.7 highly supports lineages as separate species; 0.2 < gdi < 0.7
859 denotes uncertain species status, while gdi < 0.2 denotes lineages are the same
860 species.
861
862
30
20
Number of Delimited Species
Run
Initial
Recursive
10
0
0.000 0.005 0.010 0.015 0.020
Maximal Intraspecific Distance

Firefly Chan Kin Onn

Uploaded by

Copyright:

Available Formats

You might also like

Firefly Chan Kin Onn

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Firefly Chan Kin Onn

Uploaded by

Copyright:

Available Formats

bioRxiv preprint first posted online May. 9, 2019; doi: http://dx.doi.org/10.1101/632612.

The copyright holder for this preprint (which

1 Title: Phylogeography and Molecular Species Delimitation Reveal Cryptic and

7 Authors: Wan F. A. Jusoh1, Lesley Ballantyne2 & Chan Kin Onn1*

14 *Corresponding author: chankinonn@gmail.com

16 Keywords: Lampyridae; CO1; CAD; phylogenetics; biogeography; species

143 balingiana is sufficiently genetically distinct from P. malaccae to warrant specific

147 Taxon coverage and geographic sampling

171 Phylogenetic analyses

191 Species delimitation

204 bGMYC.—compared to ABGD, the Bayesian implementation of the general mixed

219 mPTP.—the multirate Poisson tree process (mPTP) is also a phylogeny-aware

269 Phylogenetic analyses

288 Genetic divergence and species delimitation

306 bGMYC.—the distribution of the coalescence to Yule branching rate ratios

332 Phylogenetic relationships

333 Phylogenetic relationships in Luciolinae were exclusively based on ~400

357 Species boundaries and systematics

369 Pteroptyx balingiana was described as a separate species from P. malaccae

401 Caveats, limitations, and future directions

430 Another species that is morphologically similar to P. malaccae and P. balingiana is P.

530 Case, J. F. (2007). Courting Behavior in a Synchronously Flashing, Aggregative

564 de Queiroz, K. (2007). Species concepts and species delimitation. Systematic

581 amplification of mitochondrial cytochrome c oxidase subunit I from diverse

591 Fraga, H. (2008). Firefly luminescence: A historical perspective and recent

612 Biological Sciences, 270(1512), 313–321.

658 Lloyd, J. E. (2002). Lampyridae. In R. A. Arnett, M. C. Thomas, P. E. Skelley, & J. H.

660 Lloyd, J. E. (2008). Fireflies (Coleoptera: Lampyridae). In Encyclopedia of

665 McDermott, F. A. (1966). Lampyridae. In W. O. Steel (Ed.), Coleopterorum catalogus

684 Prasertkul, T. (2018). Characteristics of Pteroptyx Firefly Congregations in a Human

702 Divergence. PLoS Biology, 14(12), 1–22.

788 FIGURE LEGENDS

817 SUPPORTING INFORMATION

Number of species Prior max. distance

Partition 1 10(26) 0.001

Partition 2 10(14) 0.001668

Partition 3 10(13) 0.002783

Partition 4 10(11) 0.004642

Partition 5 10(10) 0.007743

Partition 6 9(10) 0.012915

Partition 7 1(1) 0.021544

balingiana vs. bearni Borneo tener Borneo tener Borneo+PM east

malaccae vs. bearni PM vs. tener PM east vs. tener PM west

P- distance (%) 1.48–2.92 0.84–5.00 1.39–2.29 3.15–5.49

ABGD  Lump/split Lump/split Lump/split Split

mPTP  Low (0.16) Low (0.13) Low (0.67) High (1.0)

gdi  High Uncertain Uncertain High

852 more details.

857 = moderate support; PP > 1.0 = Low support.

You might also like

ABGD Lump/split Lump/split Lump/split Split

mPTP Low (0.16) Low (0.13) Low (0.67) High (1.0)

gdi High Uncertain Uncertain High