Carrion-breeding Sarcophagidae (Diptera) can be used to estimate the post-mortem interval in forensic cases. Difficulties with accurate morphological identifications at any life stage and a lack of documented thermobiological profiles have limited their current usefulness. The molecular-based approach of DNA barcoding, which utilises a 648-bp fragment of the mitochondrial cytochrome oxidase subunit I gene, was evaluated in a pilot study for discrimination between 16 Australian sarcophagids. The current study comprehensively evaluated barcoding for a larger taxon set of 588 Australian sarcophagids. In total, 39 of the 84 known Australian species were represented by 580 specimens, which includes 92% of potentially forensically important species. A further eight specimens could not be identified, but were included nonetheless as six unidentifiable taxa. A neighbour-joining tree was generated and nucleotide sequence divergences were calculated. All species except Sarcophaga (Fergusonimyia) bancroftorum, known for high morphological variability, were resolved as monophyletic (99.2% of cases), with bootstrap support of 100. Excluding S. bancroftorum, the mean intraspecific and interspecific variation ranged from 1.12% and 2.81–11.23%, respectively, allowing for species discrimination. DNA barcoding was therefore validated as a suitable method for molecular identification of Australian Sarcophagidae, which will aid in the implementation of this fauna in forensic entomology.

Introduction	Go To >> Top Introduction Materials and methods Results and discussion Conclusions Acknowledgements References

A range of insects present on a corpse can be used as evidence in forensic investigations to estimate the post-mortem interval (PMI). Commonly, estimation of PMI using insect evidence requires accurate species identification, with subsequent examination of thermobiological profiles to determine age (Catts 1992; Catts and Goff 1992; Amendt et al. 2004). For accuracy, forensic entomologists preferentially use evidence from initial corpse colonisers, such as flesh flies (Diptera : Sarcophagidae) and blow flies (Diptera : Calliphoridae) (Amendt et al. 2004). Despite the prospective use of sarcophagids in forensic investigations, their use to date has been overshadowed by calliphorids. This is due to the difficulties of morphological species-level identification at any life stage of flesh flies, and a lack of documented thermobiological profiles of these insects.

Adult sarcophagids can be easily identified at the family level, as most species share the characteristic features of longitudinal stripes on the thorax and a tessellated/chequered abdominal pattern. However, species-level identification is difficult and requires examination of subtle morphological variation of bristle placement and length, hair colouration, body pigmentation and genitalic structure (Shewell 1987; Pape 1996). Considering this, molecular-based approaches for species identifications have been proposed to eliminate issues with identifications based exclusively on taxonomy (Wells et al. 2001; Zehner et al. 2004).

DNA barcoding is now a commonly accepted method for molecular species identification, utilising a 648-bp fragment from the 5′ end of the mitochondrial COI gene. Numerous studies have evaluated the effectiveness of barcoding, with the approach shown to be unreliable for some Diptera (Meier et al. 2006; Whitworth et al. 2007) but also proven successful for many groups of invertebrates, such as springtails (Collembola) (Hogg and Hebert 2004), butterflies (Lepidoptera) (Hebert et al. 2004), mayflies (Ephemeroptera) (Ball et al. 2005), black flies (Diptera : Simuliidae) (Rivera and Currie 2009), scuttle flies (Diptera : Phoridae) (Boehme et al. 2010) and blow flies (Nelson et al. 2007), as well as some vertebrates (e.g. Ward et al. 2005; Tavares and Baker 2008).

In a previous pilot study, 16 species of Australian Sarcophagidae were successfully resolved using DNA barcoding (Meiklejohn et al. 2011). The purpose of this initial study was to test the principle that the barcode region could distinguish between members of this important fauna. The aim of the present study was to substantially increase the earlier level of sampling. We sampled 588 sarcophagid specimens, including representatives from all Australian states and territories. This sample comprised 39 of the 84 known Australian species, represented by 580 specimens, and includes ~92% of the potentially forensically important species, mainly of the genus Sarcophaga Meigen. The remaining eight specimens that we collected could not be reliably identified, but were included nonetheless as six unidentifiable taxa. It is hoped that the results of this study will assist with the implementation of Australian flesh flies in forensic investigations.

Materials and methods	Go To >> Top Introduction Materials and methods Results and discussion Conclusions Acknowledgements References

Trapping at decayed meat baits (comprising sheep’s liver and kangaroo mince), hand netting and the ‘hill-topping’ technique of collecting from leks (Blackith and Blackith 1992), were all employed to collect adult sarcophagid specimens across Australia (Table S1 in Supplementary Material). All specimens were collected directly into absolute ethanol and stored at 4°C in the Diptera Collection in the School of Biological Sciences, University of Wollongong. Morphological species identifications were carried out by KAM for each specimen using the taxonomic keys for the Australian flesh flies (Lopes 1954, 1959). To confirm species identifications, the terminalia of each specimen were examined, which required dissections of some male specimens.

Two legs from each adult sarcophagid specimen were used as tissue for total genomic DNA extractions using a previously published protocol (Aljanabi and Martinez 1997). The DNA was resuspended in 50 μL of fresh TE solution (1 mm Tris–HCl (pH 8), 0.1 mm EDTA) and subsequently stored at 4°C. The 648-bp COI barcoding region was amplified and sequenced using the protocol outlined by Meiklejohn et al. (2011).

Sequence electropherograms were edited using ChromasPro ver. 1.33 (Technelysium, Tewantin, Qld, Australia) (available online at www.technelysium.com.au/ChromasPro.html). To confirm that the COI gene had been amplified, each sequence was submitted to both the Barcoding of Life Database Management and Analysis System (BOLD) (available online at www.boldsystems.org) and the Basic Local Alignment Search Tool (BLAST) (National Center for Biotechnology Information, Bethesda, MD, USA) (available online at http://blast.ncbi.nlm.nih.gov/Blast.cgi). All nucleotide sequences were translated into amino acid sequences using the program EMBOSS Transeq (available online at www.ebi.ac.uk/Tools/emboss/transeq/index.html) to determine the correct reading frame. ClustalW within MEGA ver. 4 was used to align all mitochondrial gene sequences (Tamura et al. 2007). All sequences were entered into BOLD, where storage and preliminary barcoding analyses were performed.

To obtain a visual representation of the divergence between specimens, a bootstrap (2000 replicates) neighbour-joining (NJ) analysis was performed using the program PAUP* ver. 4.0b10 (Swofford 2001). A semistrict consensus of the 2000 NJ bootstrap trees was generated, which retained only those relationships that had occurred on >50% of the trees. The two ‘unknown’ miltogrammine specimens included in the taxon set (KM059 and KM837), along with three blowfly (Diptera : Calliphoridae) species (Calliphora augur, Chrysomya rufifacies and Lucilia cuprina), were used as the outgroup sequences. To quantitatively evaluate DNA barcoding for the Australian Sarcophagidae, nucleotide sequence divergences were calculated using the Kimura-2-parameter (K2P) distance model, available within PAUP* (Kimura 1980).

Results and discussion	Go To >> Top Introduction Materials and methods Results and discussion Conclusions Acknowledgements References

In the present study, 588 specimens were collected from across Australia and 39 species of Sarcophagidae were identified. Importantly, some of the taxa used in the present study were from Meiklejohn et al. (2011); some of these were also misidentified. These identifications have been corrected in the current manuscript and appear with the correct species identification and same unique voucher code as in Meiklejohn et al. (2011) (denoted by ⁺ at the specimen voucher code: Table S1 in Supplementary Material).

Difficulties were encountered in accurately identifying some of the 588 specimens using the available taxonomic keys, as these keys facilitate the identification of only 54 of the possible 84 Australian species. Given that reference barcode sequences for all Australian sarcophagids are not available, correct species identifications of all specimens are vital for subsequent evaluation of the barcoding approach. To assist with the identification of specimens whose identity was uncertain, sarcophagid specialist Associate Professor Thomas Pape (Natural History Museum of Denmark) was consulted. Photographs were taken of the lateral, dorsal and head profiles, along with detailed images of terminalia of each specimen. Most of these specimens were confidently identified; however, eight female specimens could not be accurately classified to species. These specimens were nonetheless still included in the taxon set, and are represented as ‘unknown’, with some of the ‘unknown’ species comprising multiple individuals that were morphologically identical to one another. These specimens collectively represent six unidentifable taxa: Miltogramma Unknown A, Protomiltogramma Unknown A, and Sarcophaga Unknown A–D. To further assist with identifications, each ‘unknown’ sequence was submitted to BOLD and NCBI; however, no conclusive matches were obtained. We cannot confidently associate the unknown Sarcophaga species with a particular subgenus, and it is possible that these could represent new species, given that no extensive work on the Australian fauna has been carried out since the 1950s.

As complete morphological species descriptions are not available for 40% of female Australian Sarcophagidae, it could be argued that a male-only taxon set should be used to evaluate DNA barcoding for this fauna, as these are the only specimens that can be reliably identified. However, most species in the literature that lack complete descriptions of females are not likely to be carrion breeders. As this study was aimed at probable carrion-breeding Australian sarcophagids, morphological identifications of females in this study can be regarded as reliable. The specimen composition used in this study was 53% female and 47% male.

Of the 39 known and six ‘unknown’ species in the current taxon set, 33 species are represented by both sexes, three species by males only and eight species by females alone. If this study were based solely on male specimens, nearly 18% of the species diversity would be missed. Similar results have been documented by Ekrem et al. (2010) in chironomids (Diptera : Chironomidae), where females are also considerably more difficult to identify than males. In their study, 304 of 402 specimens collected were males, while 27% of species were represented only by females. Complete morphological descriptions of females for difficult species might be obtained in the future, by means of associating them with male specimens of the same species, through a combination of barcoding and further morphological work (Yeates et al. 2011).

To minimise the possibility of amplifying nuclear pseudogenes, taxon-specific primers were used for amplifications and only strongly amplified products were sequenced (Song et al. 2008; Moulton et al. 2010). Further evidence that the barcoding sequences obtained were of mitochondrial origin came from the observation that they did not contain base ambiguities or premature stop codons upon translation.

To additionally validate the barcoding approach, attempts were made to obtain reference barcode sequences for all 84 Australian sarcophagids, from pinned museum specimens dating as far back as the 1920s. Both Chelex^® (BioRad, Gladesville, NSW, Australia) and the Qiagen DNeasy Blood and Tissue kit (Qiagen, Doncaster, Victoria, Australia) extraction methods were trialled, using only one leg in each extraction to retain the integrity of the specimens. Only small ~55-bp products of COI could be amplified using the primer combination of LCO1490-L (5′-GGTCWACWAATCATAAAGATATTGG-3′) and mtD5 (5′-TGTTCCTACTATTCCGGCTCA-3′) from the DNA extracted from the pinned specimens, consistent with high levels of DNA degradation. Direct sequencing of these products failed. As this study did not exhaustively examine all methods for DNA extraction and sequencing from pinned museum specimens, future studies should examine a broad range of extraction techniques in order to obtain complete reference sequences for the Australian Sarcophagidae.

This study was aimed at comprehensively evaluating the barcoding approach for species identification of forensically important Australian Sarcophagidae. Two early methods employed for evaluating DNA barcoding for species-level resolution include: generation of a NJ tree based on K2P distances, and calculation of the intraspecific (within-species) and interspecific (between-species) sequence variation. More recent approaches have focussed on the principles of population genetics to assess barcoding (Matz and Nielsen 2005; Nielsen and Matz 2006; Pons et al. 2006). However, to allow for direct comparison with the pilot study that evaluated barcoding for the Australian Sarcophagidae (Meiklejohn et al. 2011), we have employed the two earlier approaches to evaluate barcoding in the current comprehensive study.

In successful barcoding studies, specimens morphologically identified as the same species should be resolved as a single monophyletic group within the NJ tree (Hebert et al. 2003a). Australian sarcophagid specimens that were identified as the same species according to their morphology were nearly always resolved as a single monophyletic group on the NJ tree (99.2% of cases) (Fig. 1). Sarcophaga (Fergusonimyia) bancroftorum, which appears to be the most morphologically variable Australian sarcophagid, was not monophyletic; rather, it resolved as five distinct clusters (Fig. 1). This polyphyly was not surprising given that some morphological variation was noted between the two male clusters: difference between the presence and absence of setulae on the propleuron and shape of the juxta (male terminalia). Specimens identified as the same ‘unknown’ species were also resolved separately as monophyletic on the NJ tree. Examination of the tree revealed large sequence divergences between most of the monophyletic groups, as they are separated by long branches. The monophyly of most species groups was well supported, having bootstrap values of 100 (Fig. 1). However, the species clusters of Sarcophaga (Sarcorohdendorfia) omikron and Sarcophaga (Sarcorohdendorfia) spinigera both had bootstrap support of 88, as they had one specimen more divergent than the others (Fig. 1). It was plausible that the divergent sequences, KM311 (S. omikron) and KM260 (S. spinigera) may have been from a nuclear copy of COI. However, neither sequence, when rechecked, contained premature stop codons or indels. Despite this, an additional NJ tree was generated with the removal of the two divergent specimens (KM311 and KM260). In this tree, the monophyly of S. omikron and S. spinigera was supported with bootstrap values of 100 for each taxon.

Fig. 1.  Neighbour-joining (NJ) tree of Kimura-2-parameter (K2P) distances for 623 cytochrome oxidase subunit I (COI) gene sequences from Sarcophagidae: 588 from Australian specimens and 35 from specimens collected outside Australia obtained from GenBank (denoted by an asterisk). GENERA and subgenera are given on the right-hand side: white bar at top indicates Miltogramminae, while black bars represent Sarcophaginae. Numbers given on main branches refer to bootstrap proportions among 2000 bootstrap replicates >50% (internal monophyletic bootstrap values not shown). Morphological species identifications are given for all specimens, with voucher ID and GenBank accession number given for Australian and international specimens, respectively. Outgroups consist of two unidentified species of Miltogramminae (KM059 and KM837) and three species of Calliphoridae (Calliphora augur, Chrysomya rufifacies and Lucilia cuprina). Evolutionary distance divergence scale bar, 0.1.

To assess the effect of geographic variation on the robustness of the barcoding approach, specimens were collected from various locations across Australia. For some species, such as Sarcophaga (Sarcorohdendorfia) praedatrix, Sarcophaga (Parasarcophaga) taenionota and Sarcophaga (Taylorimyia) aurifrons, monophyletic groups were resolved even with the inclusion of over 50 specimens obtained from multiple geographically isolated populations, collected over a 3500-km range. To further test population effects on species resolution, sequences from species represented in the Australian sarcophagid taxon set (but collected outside Australia), were downloaded from BOLD (accessed on 26 August 2011). In total, 35 sequences from seven species were obtained: Sarcophaga africa, S. crassipalpis, S. dux, S. misera, S. peregrina, S. ruficornis and S. taenionota. These sequences were added to the 588 Australian sequences upon initial generation of the NJ bootstrapped tree, and are denoted by an asterisk in Fig. 1. In nearly every case (33 of the 35 sequences), the international sequences were recovered in a monophyletic grouping with their Australian conspecifics (Fig. 1). In the remaining two cases, an S. dux (AY879255) specimen was recovered with Sarcophaga (Liosarcophaga) kohla, whereas one S. peregrina (EU815030) specimen grouped among the outgroup sequences. Given how difficult these species are to identify, it is likely that these specimens were misidentified, or mislabelled or the sequence contaminated by the depositor.

Calculation of the percentage divergence between sequences is used to quantitatively evaluate the success of DNA barcoding. For successful species-level resolution using the barcoding approach, interspecific genetic variation exceeds that of intraspecific variation (Hebert et al. 2003a, 2003b). This was the case for most of the Australian Sarcophagidae examined in this study, which we describe more fully in the following section. A similar result to this was obtained when the barcoding approach was previously evaluated for Australian sarcophagids (Meiklejohn et al. 2011).

The mean intraspecific variation for the Australian Sarcophagidae used in this study, excluding S. bancroftorum, ranged from 1.12% (Table 1). For 33 of the 36 species, the mean intraspecific variation was lower than 1%. The mean intraspecific variation for the nine specimens morphologically identified as S. bancroftorum was 7.67% (Table 1), which would indicate that these specimens are not a single species. Interestingly, the mean intraspecific variation of S. bancroftorum clusters KM589+KM590+KM813, KM886+KM887 and KM691+KM822 was 0.487%, 0% and 0%, respectively. These results correllate with the separation of these specimens in the NJ tree, and, as such, it is plausible that specimens identified as S. bancroftorum may represent multiple distinct species.

Table 1.  Intraspecific percentage divergences for Australian Sarcophagidae Summary of the intraspecific percentage genetic divergences (using K2P model) among 36 of the Australian sarcophagid species studied. Species that were represented by only one specimen are not shown. Symbols given at species’ names indicate biology: * denotes that the species is of potential forensic importance, and # denotes that the species is a parasitoid. For species lacking a symbol, the biology is not documented. Additionally, ^ denotes that the species has been documented in Australian forensic cases (JFW, unpublished data)

Some studies have documented that intraspecific variations can be grossly underestimated by the inclusion of numerous specimens from a single-species population (Meier et al. 2006). Given that the taxon set in this study included specimens from a range of geographical populations across Australia, it is likely that the calculated intraspecific variation has not been biased in this way. For species for which international sequences are available in BOLD, the intraspecific variation was recalculated with the inclusion of these sequences (Table 2). The intraspecific variation for all species, except S. peregrina, was lower than 2.6%. Given that one international specimen of S. peregrina (EU815030) did not cluster with its conspecifics, the higher variation of 3.82% was due to the inclusion of this specimen in the calculations (1.33% when EU815030 was removed). Low intraspecific barcode variation was previously documented in Sarcophaga (Liopygia) argyrostoma (0.6–2.6%) and for a range of sarcophagids from both Europe and the USA (<1%) (Wells et al. 2001; Zehner et al. 2004; Draber-Monko et al. 2009). Overall, the findings indicate that, even with a comprehensive sample of the Australian Sarcophagidae, this fauna mainly has low intraspecific variation, making barcoding an appropriate molecular identification method.

Table 2.  Intraspecific percentage divergences among Australian Sarcophagidae, including international conspecifics Intraspecific percentage genetic divergences (using K2P model) of seven Australian Sarcophagidae with the inclusion of international sequences

The interspecific variation was calculated for Australian sarcophagid species that were clustered closely together on the NJ tree (Table 3). Specimens of S. megafilosia and S. meiofilosia had an interspecific variation of 2.81%. It is known that both of these species are parasitoids of the marine snail Littoraria filosa, where S. megafilosia only parasitises snails with shell lengths ≥10 mm and S. meiofilosia parasitises snails with shells 4–10 mm long (Pape et al. 2000). Given the similar biology, restricted Queensland distribution and the short branch lengths linking these flesh fly species, the lower-than-expected interspecific variation between them suggests that they have only recently diverged or are not separate species. For the remaining species that were compared, the mean interspecific variation was 3.75–11.23%. This range is in accordance with other studies that have used the COI gene for sarcophagid identification (Wells et al. 2001; Zehner et al. 2004).

Table 3.  Interspecific percentage divergences between Australian Sarcophagidae Interspecific percentage genetic divergences (using K2P model) between those Australian sarcophagids that clustered closely together on the NJ tree

Meier and colleagues (Meier et al. 2008) have cautioned against using mean interspecific divergences as a method for species discrimination in Diptera. They propose that, with increases in specimen numbers, mean interspecific variations become inaccurately inflated. Accordingly, they suggest that the only correct reflection of species variation is from the minimum interspecific variation. With the exception of the S. megafilosia and S. meiofilosia comparison discussed above, the smallest mean interspecific divergence found was for S. crassipalpis versus S. ruficornis (3.75%), which is not greatly different from the minimum interspecific divergence (3.61%) for specimens of this pair (Table 3). Similar results are seen for all species pairs, suggesting that the use of mean interspecific variation does not inflate the gap between intraspecific and interspecific variation for Australian sarcophagids.

We emphasise that the NJ tree presented in this manuscript does not attempt to resolve any relationships between the subfamilies, genera, subgenera and species of the Australian Sarcophagidae. Bootstrap support at higher-level nodes is not given on the NJ tree as it was <50, indicating a lack of confidence in the relationships depicted. Given that the main application of the COI barcoding region is for molecular species-level identifications, we were also unable to reliably associate the six unknown Sarcophaga species with a particular subgenus. More substantial work is needed to identify these unknown species, which might be achieved by obtaining sequence data from genes in addition to COI, and employing a range of phylogenetic methods to investigate their relationships.

Conclusions	Go To >> Top Introduction Materials and methods Results and discussion Conclusions Acknowledgements References

To date, forensic entomologists have not capitalised on using flesh fly evidence for estimating the PMI, due to the difficulties inherent in morphological identifications, along with a lack of thermobiological data. This study focussed on comprehensively evaluating DNA barcoding as a molecular approach for the identification of 588 specimens of the Australian Sarcophagidae. Examination of percentage genetic divergences and a NJ tree were used to test barcoding in this study, with the results indicating that it is an effective approach for the accurate species resolution of this fauna, in line with our previous study focussed on the east coast (Meiklejohn et al. 2011). We advocate, however, that future barcoding studies on sarcophagids should compare multiple methods for evaluating barcoding, including population genetic approaches. Following the accurate identification of a flesh fly specimen, the relevant thermobiological profile should be examined to determine the specimen’s age. Currently, such profiles are not available for most sarcophagids. Future studies should therefore also focus on documenting flesh fly thermobiology, in order to fully facilitate the use of these flies as effective tools in forensic entomology.

Acknowledgements	Go To >> Top Introduction Materials and methods Results and discussion Conclusions Acknowledgements References

We gratefully acknowledge the Australian National Insect Collection (CSIRO), Australian Museum, Queensland Museum, Melanie Archer, Bryan Cantrell, Kelly George, Bryan Lessard, Christopher Manchester, Steve and Ruth McKillup, Lisa Mingari and Leigh Nelson for providing specimens, and David Boyd for his computing assistance. We thank Thomas Pape for his assistance with specimen identifications and the Australian Biological Resources Study, the Australian Research Council, the Australian Federal Police, the NSW Police Force and the Taxonomic Research Information Network (Emerging Priorities Program) for their financial support of this study.

References	Go To >>\ Top Introduction Materials and methods Results and discussion Conclusions Acknowledgements References