Comparison of Matrix-Assisted Laser Desorption Ionization–Time of Flight Mass Spectrometry and Molecular Biology Techniques for Identification of Culicoides (Diptera: Ceratopogonidae) Biting Midges in Senegal

Abstract

Biting midges of the genus Culicoides are implicated as vectors for a wide variety of pathogens. The morphological identification of these arthropods may be difficult because of a lack of detailed investigation of taxonomy for this species in Africa. However, matrix-assisted laser desorption ionization−time of flight mass spectrometry (MALDI-TOF MS) profiling is efficient for arthropod identification at the species level. This study established a spectrum database of Culicoides spp. from Senegal using MALDI-TOF. Identification of Culicoides insects to the species level before mass spectrometry was performed on the basis of morphological characters. MALDI-TOF MS reference spectra were determined for 437 field-caught Culicoides of 10 species. The protein profiles of all tested Culicoides revealed several peaks with mass ranges of 2 to 20 kDa. In a validation study, 72 Culicoides specimens in the target species were correctly identified at the species level with a similarity of 95 to 99.9%. Four Culicoides protein profiles were misidentified. Nevertheless, six SuperSpectra (C. imicola, C. enderleini, C. oxystoma, C. kingi, C. magnus, and C. fulvithorax) were created. Abdomens of midges were used to amplify and sequence a portion of the mitochondrial cytochrome oxidase I gene (COI). The results obtained using the MALDI-TOF MS method were consistent with the morphological identification and similar to the genetic identification. Protein profiling using MALDI-TOF is an efficient approach for the identification of Culicoides spp., and it is economically advantageous for approaches that require detailed and quantitative information of vector species that are collected in field. The database of African Culicoides MS spectra created is the first database in Africa. The COI sequences of five Culicoides species that were previously noncharacterized using molecular methods were deposited in GenBank.

INTRODUCTION

Biting midges of the genus Culicoides are among the smallest hematophagous insects, and they can be a nuisance to humans and animals (1). These insects are implicated as vectors for a wide variety of viruses, such as bluetongue virus (2), African horse sickness virus (3), epizootic hemorrhagic disease virus (4), the Schmallenberg virus that recently emerged in Europe (5–7), Toggenburg orbivirus (8), and the Oropouche virus (9), which is the only human virus transmitted by Culicoides. Additionally, biting midges are involved in the transmission of various human filarial parasites, such as Mansonella perstans (10–13), Mansonella streptocerca, Mansonella ozzardi, and Mansonella rodhaini (14). Recently, Leishmania spp. and Leishmania infantum were detected in Culicoides imicola and Culicoides spp. (15).

Vector identification is an important step in the epidemiology of vector-borne diseases. Information on the major vector species may provide a clearer indication of the geographic distribution of the disease or its potential distribution, the location of danger points for high risks of contact with the vector, and access to alternative tools for the study of the natural cycles of parasites. A comprehensive key to all of the species within a genus in any particular faunal region is a useful, primary tool for taxonomists (16). However, identification of the Culicoides species is challenging because the morphological identification of these tiny insects to the species level is very difficult in many cases (17, 18). Identification requires time-consuming microscopic analyses of slide-mounted insect preparations (19). The scarcity of data on Culicoides taxonomy in Western Africa also complicates identification. Several PCR-based tests were developed for the identification of Culicoides spp., and these tests target the mitochondrial cytochrome oxidase I gene (COI) or the rRNA genes, internal transcribed spacer 1 (ITS-1) or 2 (20–26).

Matrix-assisted laser desorption ionization−time of flight mass spectrometry (MALDI-TOF MS) was tested for 2 decades as an alternative method for microorganism identification. This approach has developed for the high-throughput, accurate, and reproducible identification of clinically relevant microorganisms (bacteria, yeasts, and filamentous fungi) at a low cost with minimal preparation time (27–32).

Additionally, this proteomic approach has been used for the identification of metazoans, such as fish species (33), plants (lentil varieties) (34), ticks (35, 36), and insects, such as Drosophila fruit flies (37, 38) and mosquitoes (39, 40).

Recently, similar studies demonstrated the suitability of MALDI-TOF MS analysis for the characterization of Culicoides flies (3, 41, 42) in Europe. Therefore, molecular and MALDI-TOF approaches are promising techniques for the elucidation of relationships within species and between closely related species.

The present study investigated the ability of MALDI-TOF MS to consistently characterize biting midges of the Culicoides genus in Africa and established a baseline taxonomy based on mass spectrometry.

MATERIALS AND METHODS

Insect collection.

Insects were collected using a modified CDC light trap (43). Briefly, tulle net pockets were replaced by pocket sails of terylene, which traps tiny insects. Specimens were collected in three regions of Senegal (Fig. 1): Kédougou, a vast region of southeastern Senegal surrounded by large hills; Sine-Saloum, a region in the northwestern part of the Gambia basin; and Lower Casamance (Oussouye department), a region in the southwestern part of Senegal (Table 1). Traps were placed in different habitats, which resulted in important variations in their qualitative and quantitative performance. The majority of captured specimens were female. Irregularities in catching made it very difficult to assess seasonal variations.

FIG 1.

Map of the villages where Culicoides spp. were collected during the study. In the Sine-Saloum, inspected villages were Dielmo, Ndiop, Toubacouta, Sourou, and Nema Bah; in Casamance, Mlomp, Kagnoute, and Elinkine were inspected; and in Kédougou, Bandafassi, Mbaning, Ibel, and Tiabéji were inspected.

TABLE 1.

Villages where Culicoides spp. were collected during this study

Study site	Climate vegetation	Annual precipitation (mm)	Village	Coordinates
Sine-Saloum	Sudanian wooded savannah	939	Dielmo	13°43′N, 16°24′W
			Ndiop	13°41′N, 16°23′W
			Toubacouta	13°47′N, 16°28′W
			Sourou	13°45′N, 16°28′W
			Nema Bah	13°44′N, 16°28′W
Basse Casamance	Sub-Guinean/primary and secondary gallery forest	1,432	Mlomp	12°32′N, 16°34′W
			Kagnoute	12°33′N, 13°34′W
			Elinkine	12°30′N, 16°39′W
Kédougou	Sudano-guinean/woodland,wooded savannah	1,189	Bandafassi	12°32′N, 12°18′W
			Ibel	12°30′N, 12°22′W
			Tiabéji	12°38′N, 12°27′W
			Mbaning	12°31′N, 12°27′W

Morphological identification.

Culicoides spp. were separated from other insects using a stereomicroscope and sorted on the basis of wing morphology into the groups of C. schultzei, C. imicola, C. magnus, C. milnei, and other Culicoides spp. (44, 45). Species were identified using microscopic analyses of wing patterns and the observation of different body parts (head, legs, and wing) (44–46). Insects were stored individually at −80°C after identification for further DNA and mass spectrometry analyses.

Sample preparation for MALDI-TOF MS.

Abdomens of Culicoides specimens were removed for further molecular studies after dissection using a stereomicroscope, and the remaining parts were individually transferred to 0.2-ml tubes. The remaining parts were triturated in 10 μl of 10% formic acid using a manual homogenizer with disposable inoculation loops. One microliter of the homogenate was spotted in duplicate onto a steel target plate. Additionally, 1 μl of alpha-cyano-4-hydroxycinnamic acid (CHCA) matrix was added directly to the spots and air dried at room temperature. Control experiments and calibration were performed using an Escherichia coli reference strain (LyfoCults E. coli ATCC 8739; bioMérieux). Plates were submitted for MALDI-TOF MS analysis after air drying.

MALDI-TOF MS analysis.

Protein profiles were obtained using MALDI-TOF Vitek mass spectrometry (Vitek MS RUO system; bioMérieux, Marcy l'Etoile, France) at the Principal Hospital of Dakar. The resulting profiles were analyzed using SARAMIS Premium software, v.4.0.0.14. Potential species-specific biomarker mass patterns, called SuperSpectra, were calculated using the SARAMIS SuperSpectra tool. Therefore, peak lists of representative isolates were imported into SARAMIS Premium software. Peak lists were trimmed to a mass range of 2 to 20 kDa, and peaks with a relative intensity of <5% were removed. Peak lists were binned, and average masses were calculated using the SARAMIS SuperSpectra tool.

We used the SARAMIS Premium software package to create an MS species-based dendrogram. The dendrogram was based on the entire spectra of Culicoides spp. that were collected during this study. The threshold for identification was set at 75% of biomarker matches based on the reference data set in accordance with SARAMIS Premium user guidelines.

A binary mass list was calculated with an error of 800 ppm, and the SARAMIS single-link clustering algorithm was applied.

Extraction of DNA and PCR amplification.

Mechanical separation of the abdomens of biting midges was performed under a stereomicroscope using fine forceps, disposable petri dishes, needles, and razor blades. Disposable instruments were washed in 70% ethanol between specimens of the same species and discarded after each species was treated.

The abdomen of each biting midge was placed in a 1.5-ml tube that contained 100 μl of lysis buffer G2 (30 mM Tris Cl, 30 mM EDTA, 5% Tween 20, 0.5% Triton X-100, and 800 mM guanidine hydrochloride [GuHCl]) and 20 μl of proteinase K (activity equal to 600 milliabsorbance units [mAU]/ml solution or 40 mAU/mg of protein). The homogenate was digested in a heating block for 1 h or overnight at 56°C. Total abdomen DNA was extracted using EZ1 (Qiagen, Courtaboef, France), according to the manufacturer's instructions (EZ1 DNA tissue kit; Qiagen). DNA was eluted in 100 μl and stored at −20°C until further use.

Molecular analyses of biting midges were conducted using primers C1-J-1718-F and C1-N-2191-R (47), which amplify approximately 520 bp of the COI gene. PCR consisted of an initial heating phase at 95°C for 15 min, followed by 40 cycles of denaturation (95°C for 30 s), annealing (50°C for 30 s), and elongation (72°C for 1 min) and a final elongation step (72°C for 10 min). Negative controls were used to detect possible contamination, and positive controls ensured appropriate amplification conditions. All amplifications were made in a total reaction volume of 25 μl, which included 2.5 μl of PCR 10× buffer, 0.75 μl of MgCl₂ (25 mM), 2.5 μl of deoxynucleoside triphosphate (dNTP) (20 mM), 0.5 μl of each primer, 0.1 μl of HotStarTaq DNA polymerase (5 units/μl), 15.15 μl of distilled water, and 5.0 μl of the DNA template. PCR products were loaded onto a 1.5% agarose gel stained with GelRed and scanned to assess amplifications.

Sequencing and molecular data analysis.

PCR products that displayed positive bands on gels were sequenced using a forward primer and a reverse primer and the BigDye Terminator v.1.1 kit (Life Technologies, Carlsbad, CA, USA). Products were purified and analyzed using an ABI PRISM 3130X/genetic analyzer (Applied Biosystems; Life Technologies). Sequences were analyzed using BLAST sequencing and compared with the available Culicoides sequences in GenBank.

Sequence editing and analyses were performed using BioEdit sequence alignment editor (v.7.0.5.3) (48).

A phylogenetic tree was created using the Bayesian phylogenetic analysis (49) in TOPALi 2.5 software (Biomathematics and Statistics Scotland, Edinburgh, United Kingdom) within the integrated MrBayes application (http://mrbayes.csit.fsu.edu/) using the sequences that were obtained during this study and additional COI sequences of Culicoides imicola, Culicoides oxystoma, Culicoides schultzei group spp., Culicoides obsoletus, Culicoides sonorensis, Culicoides loxodontis, and Culicoides pseudopallidipennis from GenBank (Table 2). The number of identical sequences is shown next to the species name, but only unique sequences were used in the tree.

TABLE 2.

Culicoides species COI sequences used for phylogenetic analyses^a

Species	GenBank accession no.	Geographical origin
C. enderleini	KF186430, KF186429, KF186431	Madagascar
C. enderleini	HQ447066, HQ447065	Reunion Island
C. schultzei group	JN545050, JN545051	Israel
C. nevilli	KF186428	Madagascar
C. oxystoma	JN545048, JN545047	Israel
C. oxystoma	KF528692, KF528693	China
C. oxystoma	AB360985	Japan
C. sonorensis	JF870510	United States
C. obsoletus	JA620142	Sweden
C. loxodontis	AF069235, AF069234	South Africa
C. pseudopallidipennis	AY286329	Cote d'Ivoire
C. imicola	AF080597	Israel
C. imicola	HQ447069	Reunion Island
C. imicola	AF069233	South Africa
C. imicola	AF080539	Spain
C. imicola	HQ824456	France
C. imicola	AJ549415, AF079977	Portugal
C. imicola	EU189057	Algeria

^aSequences from GenBank for 26 Culicoides spp. are shown.

Nucleotide sequence accession numbers.

All COI sequences obtained were deposited in GenBank under the following accession numbers: KJ833678 to KJ833717 and KP025627 to KP025651.

RESULTS

Morphological identification.

Morphological analyses of 437 individuals from 3 collection sites revealed the presence of 10 species that belonged to the following groups: C. imicola (C. imicola), C. schultzei (C. enderleini, C. kingi, C. oxystoma and C. nevilli), C. milnei (C. milnei, C. wansoni, and C. moreli), C. magnus, and C. fulvithorax (Fig. 2). Several damaged and unidentifiable samples were excluded from the study. Wing characterization is unique for each species, but slight variations in wing patterns between individuals within a species were frequently observed. C. moreli and C. milnei were collected only at the Sine-Saloum site, and C. kingi was collected more frequently at the Kédougou site. The other species were abundant and equally present at the 3 sampled sites.

FIG 2.

Photographs of wings of slide-mounted representatives of Culicoides specimens viewed under a light microscope. The same individuals were used for molecular and MALDI-TOF identification.

Molecular characterization and phylogenetic analysis of Culicoides species based on the COI gene.

A total of 150 morphologically identified randomly selected Culicoides specimens representing all 10 species were subjected to specific PCR using the COI primers. A total of 68 randomly selected amplicons that represented all studied species were sequenced. BLAST searches (http://blast.ncbi.nlm.nih.gov/Blast.cgi) of C. imicola, C. enderleini, C. oxystoma, C. nevilli, and C. kingi were almost identical (98 to 100%) to the sequences that were already deposited in GenBank (C. imicola, C. enderleini, C. oxystoma, C. nevilli, and C. schultzei groups). Sequences of C. milnei, C. wansoni, C. moreli, C. magnus, and C. fulvithorax were not available in GenBank before our study, and the closest identity of the obtained sequences of these species was approximately 80 to 89% with different Diptera species. Altogether, the identification of African biting midges using molecular methods was hardly possible because of the insufficient number of deposited sequences of the COI gene in GenBank.

The phylogenetic tree (Fig. 3) revealed that the 31 individual midges that were identified morphologically as C. enderleini split into 3 different well-defined and well-supported genetic groups with high homogeneity inside of the groups. One genetic group consisted of midges that were collected in Dielmo (Sine-Saloum region), and the C. enderleini sequences obtained from GenBank. The second genetic group consisted of midges that were collected from Dielmo and Kédougou. The third genetic group consisted of midges collected from Dielmo. The sequences of C. imicola obtained were closely related to the published sequences of C. imicola worldwide. The COI gene sequences from the C. wansoni specimens were grouped with C. milnei in a separate branch. C. oxystoma from Kédougou formed a sister clade to C. oxystoma from Japan, China, and Israel. C. kingi from Kédougou and Dielmo were almost identical to the deposited sequences of the C. schultzei group from Israel (Fig. 3).

FIG 3.

A consensus phylogenetic tree showing the relationships of the studied species of Culicoides based on COI sequence comparisons. GenBank accession numbers are indicated. The sequences were aligned using ClustalW, and phylogenetic inferences were obtained using Bayesian phylogenetic analysis with TOPALi 2.5 software (Biomathematics and Statistics Scotland, Edinburgh, UK) within the integrated MrBayes application using the SYM+Γ, JC, and SYM+Γ substitution models for the first, second, and third codons, respectively. Numbers at the nodes are percentages of bootstrap values obtained by repeating the analysis 100 times to generate a majority consensus tree. The final set includes 471 bp. The scale bar represents a 2% nucleotide sequence divergence. Numbers 1, 2, and 3 indicate genetic groups of C. enderleini.

MALDI-TOF MS analysis. (i) Creating a reference database of biomarker masses.

MALDI-TOF MS reference spectra were determined for 437 field-caught Culicoides biting midge species from Senegal, including C. imicola (n = 52), C. enderleini (n = 130), C. oxystoma (n = 115), C. nevilli (n = 22), C. kingi (n = 44), C. magnus (n = 22), C. wansoni (n = 23), C. milnei (n = 2), C. moreli (n = 1), and C. fulvithorax (n = 26). Protein spectral profiles of all tested Culicoides species revealed several peaks with mass ranges of 2 to 20 kDa and high signal intensities. These protein spectral profiles were very similar between specimens of the same species. Profiles of specimens belonging to these biting midge species (C. imicola, C. enderleini, C. oxystoma, C. nevilli, C. kingi, C. magnus, C. wansoni, C. milnei, C. moreli, and C. fulvithorax) are shown and ranged between 2 and 15 kDa (Fig. 4). The protein profiles of 1 to 5 specimens per species were used to compile the total mass spectra for these species. The resulting dendrogram revealed that all Culicoides specimens of the same species were clustered on distinct branches (Fig. 5). Additionally, six SuperSpectra were created for C. imicola, C. enderleini, C. oxystoma, C. kingi, C. magnus, and C. fulvithorax.

FIG 4.

MALDI-TOF Vitek mass spectra in the range of 2 to 14 kDa of 10 field-caught Culicoides insects (C. kingi, C. imicola, C. nevilli, C. moreli, C. magnus, C. wansoni, C. fulvithorax, C. enderleini, C. milnei, and C. oxystoma). Six SuperSpectra (in green), C. kingi, C. imicola, C. magnus, C. fulvithorax, C. enderleini, and C. oxystoma, were created. The most remarkable specific peaks are indicated by the arrows.

FIG 5.

Dendrogram of MALDI-TOF Vitek mass spectra in the range of 3 to 20 kDa of thoraxes from 2 to 4 individuals of 10 species of field-caught Culicoides biting midges. Ma, C. magnus; Fu, C. fulvithorax; En, C. enderleini; Ki, C. kingi; Ne, C. nevilli; Mo, C. moreli; Wa, C. wansoni; Mi, C. milnei; Im, C. imicola.

(ii) Study validation.

The accuracy of the Culicoides reference database was tested in a second step using a validation study. A total of 76 Culicoides specimens (C. imicola, n = 15; C. enderleini, n = 19; C. kingi, n = 11; C. magnus, n = 7; C. fulvithorax, n = 8; and C. oxystoma, n = 16) were blindly tested (i.e., the spectra obtained were tested using a created database). A comparison of all of these specimens in the database using Vitek MS software revealed satisfactory results (Table 3). Among the 72 specimens of biting midge species tested, the correct identification was achieved at the species level with a confidence index from 95 to 99.9% similarity. One C. enderleini and 3 C. oxystoma protein profiles were misidentified. C. enderleini was identified as C. oxystoma with an average of 99.99% similarity, and the 3 C. oxystoma were identified as C. enderleini with 98 to 99.90% similarity. Some of these Culicoides species were sequenced.

TABLE 3.

Validation of Vitek MS results for 76 Culicoides spp.

Species	No. of morphological identifications	Correspondence of COI identification to morphological identification (no./total no. [%])	Correspondence of MALDI-TOF identification to morphological identification (no./total no. [%])
C. imicola	15	15/15 (100)	15/15 (100)
C. fulvithorax	8	8/8 (100)	8/8 (100)
C. magnus	7	7/7 (100)	7/7 (100)
C. kingi	11	11/11 (100)	11/11 (100)
C. enderleini	19	18/19 (95)a	18/19 (95)
C. oxystoma	16	16/16 (100)	13/16 (69)b
Total	76	75/76 (99)	72/76 (95)

^aOne individual was morphologically identified as C. enderleini. The sequenced COI and MALDI-TOF spectra corresponded to C. oxystoma.

^bAll 16 specimens that were morphologically identified as C. oxystoma were confirmed using sequencing, but the MALDI-TOF spectra of 3 individuals were closer to those of C. enderleini than to those of other C. oxystoma.

DISCUSSION

Mass spectrometry profiling generates and identifies molecules according to mass and charge. Mass spectrometry is used for numerous applications in different areas (50). Insect identification using MALDI-TOF MS is a novel entomological tool that has been utilized in only a few studies (3, 37, 38, 40, 51), including a very recent application for the identification of cryptic Anopheles mosquito species (39). Kaufmann et al. applied this proteomic approach to characterize a hematophagous Culicoides species from Europe (41). Additionally, Steinmann et al. (42) recently evaluated this technique for the identification of Ceratopogonidae and culicid larvae. The present study used MALDI-TOF MS for identification and established the first database reference of hematophagous Culicoides in Western Africa. The database created allows for cost-effective and rapid identification of Culicoides species with high specificity. Some of the 12 species collected in this study are implicated as a vectors for protozoa and arboviruses (C. imicola, C. fulvithorax, C. kingi, and C. enderleini) (1, 15, 52–55). These data highlight the veterinary and clinical need for the species identification of Culicoides.

Our results confirm that the database of protein extract spectra obtained using MALDI-TOF is a suitable tool for the identification of Culicoides. This method was used throughout the experiments and yielded high-quality spectral profiles (Fig. 4). Kaufmann et al. (3) reported that male and female specimens had very similar spectra with a high count of identical masses, but the presence of blood had a considerable impact on the MALDI-TOF analysis because it reduced the intensity of the midges' biomarker masses. This effect was undetectable in insects 5 days after the ingestion of blood. Our study used only unfed females for MALDI-TOF identification.

Several authors evaluated other MALDI-TOF MS-based systems and reported the necessary use of a formic acid-based protein extraction using a Culicoides lysis step before matrix application (24–42). The Vitek MS system used in this study generated a low frequency of unusable spectra without formic acid, and we obtained good-quality spectra with formic acid. In general, MS provides good identification of Culicoides with satisfactory results for the species studied. MS was effective for the identification of C. imicola, C. kingi, C. fulvithorax, C. magnus, and C. enderleini, which resulted in the creation of SuperSpectra for these species. The reproducibility and accuracy of the reference spectra were tested in a blind study of 76 morphologically identified biting midges (Table 3), and a correct and unambiguous result was obtained in 72 of the 76 insects tested. Four of the insects that were misidentified by the Vitek MS species belonged to the C. schultzei group (1 C. enderleini and 3 C. oxystoma). One C. enderleini was identified by MALDI-TOF as C. oxystoma with 99.99% similarity, and 3 C. oxystoma were identified as C. enderleini with 91 to 99.99% similarity. One misidentification (Table 3) was evidently the fault of morphological identification. C. oxystoma was misidentified as C. enderleini because both sequencing and MALDI-TOF methods were concordant, and this specimen was identified as C. oxystoma. MALDI-TOF spectra were misleading in three other cases: the individuals who were identified morphologically and genetically as C. oxystoma had spectra that were closer to those of C. enderleini. All of these identifications concerned only two taxonomically related species of the C. schultzei group, C. oxystoma and C. enderleini, which resembled each other morphologically and genetically.

Comparisons of the dendrograms that were obtained using molecular biological methods and MALDI-TOF MS showed similar topologies for C. kingi, C. imicola, C. magnus, C. fulvithorax, C. oxystoma, C. nevilli, C. moreli, and C. wansoni. Few differences were shown for C. enderleini; phylogeny tree analysis revealed three clusters, but spectral analysis did not support this separation. Spectral analysis showed one cluster with similarity within the species that was >75%. This result may be explained by the few protein differences between the molecularly identified clusters, which may represent either high intraspecies variation or subspecies.

Morphological identification provides gold standard specimens for insect taxonomy. However, morphological identification can be a time-consuming procedure, and it is a very difficult task in many cases even for expert taxonomists (28) because of faint characteristics or intraspecific variability (26). Pagès et al. (56) described the existence of morphologically similar midges (cryptic species) that are genetically distinguishable. Molecular identification assays are accurate tools for species diagnosis, but precautions are required during several steps from the storage of specimens/samples to PCR amplification to ensure correct identification. Analyses of partial COI sequences confirmed previous findings (57) that this locus displays low intraspecies variation and considerable interspecies variations in Culicoides spp. (26). The molecular identification of African biting midges may be difficult because of the virtual paucity of deposed sequences; sequences for only 5 of the 10 species studied were available in GenBank. The population genetic structure of the investigated midge species might differ across ecological conditions (26). MALDI-TOF MS has emerged as an alternative technique for species identification. The technical preparation of midges included abdomen removal because host blood in female midges blurred the spectra during the first few days after the blood meal (41). This proteomic approach has been developed for use for use as a high-throughput, accurate, and reproducible identification of clinically relevant midges species at a low cost with minimal preparation time. Kaufmann et al. (41) reported that this approach may become the method of choice for the centralized, robust and high-throughput screening of midge populations in connection with the surveillance of eventually emerging midge-transmitted agents. Morphological, molecular, and protein profiling using MALDI-TOF approaches together hold promise for the elucidation of relationships within species complexes and between closely related species.

In conclusion, we report for the first time the successful identification of epidemiologically important arthropods (Culicoides spp.) using mass spectrometry that was performed entirely in Africa. The database of Culicoides spectra obtained allowed the cost-effective and reproducible identification of species without the assistance of trained entomologists. Our database contains reproducible reference SuperSpectra for 6 species of field-caught Culicoides and reference spectra for other Culicoides. This database could benefit any entomological laboratory that is equipped with a MALDI-TOF Vitek MS system. Our study also showed that Vitek MS is an efficient approach for the identification of Culicoides. This method is economically advantageous for approaches that require detailed and quantitative information on midge fauna.