Evaluation of a protocol for remote identification of mosquito vector species reveals BG-Sentinel trap as an efficient tool for Anopheles gambiae outdoor collection in Burkina Faso
© Pombi et al.; licensee BioMed Central. 2015
Received: 3 December 2014
Accepted: 4 April 2015
Published: 15 April 2015
Feasibility and costs of monitoring efforts aimed to monitor mosquito species are strictly dependent on the presence of skilled entomologists directly in the field. However, in several contexts this is not possible or easy to organize, thus limiting the possibility to obtain crucial information on presence/abundance of potential disease vectors and of new invasive species. Digital imaging approaches could be extremely useful in the frame of medical entomology to overcome this limit. This work describes a surveillance approach to collect and morphologically identify host-seeking malaria vectors based on remote transmission of digital images of specimens collected with ad hoc modified traps.
CDC light trap (CDC) and the BG-Sentinel trap (BG), both baited with BG-lure and CO2, were modified in order to have collected mosquitoes immobilized on a bi-dimensional surface. The performance of the two traps in the field was comparatively tested by Latin-square experiments in two villages of Burkina Faso under low and high mosquito densities. The efficiency of identifications based the inspection of digital images versus microscopic identifications of collected specimens was compared.
A total of 1,519 mosquitoes belonging to 16 species were collected, of which 88.5% were microscopically identified as Anopheles gambiae s.l. (mainly Anopheles coluzzii, 85.7%). During dry season BG collected 15 times more females than CDC outdoors, whereas indoors the BG collected 0.4 times less than CDC. During rainy season the ratio BG/CDC was 6.4 and 0.7 outdoors and indoors, respectively. The efficiency of digital images versus microscopic identifications of collected specimens was 97.9%, 95.6% and 81.5% for Culicidae, Anophelinae and An. gambiae s.l., respectively.
Results strongly encourage the use of BG-trap for collecting host-seeking An. gambiae particularly in the outdoor environment, providing new perspectives to the challenge of collecting this fraction of the biting population, whose epidemiological relevance is increasing due to the success of large-scale implementation of indoor malaria vector control strategies. Moreover, results show that the transmission of digital images of specimens collected by the ad hoc modified host-seeking traps efficiently allows identification of malaria vectors, thus opening the perspective to easily carry out mosquito monitoring also in the absence of entomologists directly in the field.
In recent years, the technological improvements of digital devices, data storage capacities and computer performances have radically changed the way researchers use images to collect and store information for scientific purposes. In the study of life history traits of biological organisms, such progresses have enabled the tracking of individual behaviour, fecundity or growth trajectory on a fine time scale and over long periods, leading to the development of automated tools for the analysis of digital pictures of laboratory microcosms . Moreover, semi-automated quantification of bi-dimensional traits for digital plant phenotyping has also been successfully developed . In the field of museum biological collections, digital imaging approaches have been pursued to obtain improved high-throughput workflows for digitizing and providing access to invertebrate collections, opening the field to cybertaxonomy [3-6].
However, digital approaches have been so far limited to laboratory settings and never applied to field studies in the frame of which they could be extremely useful particularly for medical and agricultural entomology surveys. In fact, the presence in the field of specialized entomologist(s) is, nowadays, a sine qua non condition for surveillance/monitoring activities aimed to quickly obtain information about the presence of insect pests or of new invasive species. This need significantly complicates the logistic of the monitoring schemes and increases their costs. These issues could be overcome by a system relying on remote transmission of digital images of collected insect samples to a reference laboratory for species identification.
The first requirement to develop such as “digital identification system” is the immobilization of collected insect specimens on a bi-dimensional surface, such as a sticky-film. Several kinds of sticky-traps have already been developed and employed in the field to passively collect mosquito vectors at different physiological stages. The collection of gravid mosquitoes with adhesive ovitraps has been exploited in several studies, most of which targeting Aedes aegypti, Aedes albopictus, Culex pipiens/quinquefasciatus and other container-breeding mosquitoes in tropical as well as in temperate areas [7-15]. The Sticky Resting Box has proved efficient for collecting resting Anopheles gambiae, both in indoor and outdoor environments in Burkina Faso . Sticky-sheets floating over breeding sites have proved effective for collecting ovipositing Anopheles females and other Culicidae in Tanzania . A wire mesh glue trap with fruits/seedpods suspended on skewers inside has been used to collect sugar-seeking An. gambiae Mali . However, none of these sticky-traps targets host-seeking mosquito females, despite the fact that these represent the most epidemiologically important fraction of vector populations.
This work reports the development of an approach for the remote monitoring of host-seeking mosquito vectors that has been tested in a highly endemic malaria area of West Africa. This approach implied the ad hoc modification of two traps routinely used to collect host-seeking mosquito females - i.e. the CDC light trap and the BG-Sentinel trap – in order to directly collect mosquitoes immobilized on a bi-dimensional surface for subsequent digital morphological identification.
A total of five 4x4 Latin-square experiments were carried out in the villages of Goden and Koubri (Ouagadougou area, Burkina Faso) during rainy (13–16 September 2011, 10–13 and 17–20 September 2012) and dry (4–10 and 16–19 April 2012) seasons in four family compounds/village. In each compound, four traps (two BG and two CDC traps) were rotated daily from two indoor and two outdoor positions, inside and outside two inhabited houses, respectively, for a total of 16 rotations/Latin square and 320 trap/days. Traps were activated for 12 hours starting at 6:00 PM. Batteries and CO2 source were replaced daily.
Mosquito identification and evaluation of digital identification protocol
The digital pictures were visually inspected in order to identify genus, species and gender of collected specimens. Digital identifications were confirmed by microscopic identifications. Anopheles gambiae s.l. specimens were isolated cutting the sticky-surfaces around them and dipping these in acetone for two minutes. Isolated mosquitoes were stored in tubes containing silica-gel for subsequent molecular species identification by PCR-SINE X protocol .
Variation in mosquito abundance between collection methods was investigated using Generalized Linear Mixed Models . In these analyses, the sampling method was fitted as a main effect, with season, village, compound, house and date fitted as random effects. Mosquito abundance data are typically overdispersed, consequently a Poisson or Negative Binomial distribution (with or without zero-inflation) were tested to model the data, choosing a Negative Binomial distribution in all cases (with fit being assessed by both Likelihood Ratio Test and Akaike Information Criterion). The predicted mean numbers of female mosquitoes of An. gambiae complex species (i.e. Anopheles coluzzii, An. gambiae s.s., Anopheles arabiensis) collected by each trapping method were obtained by extracting and exponentiating the coefficients and associated standard errors predicted from statistical models. All statistical analyses were conducted in R, v3.1.1  using the lme4  and glmmADMB  packages.
Results and discussion
This work presents an approach to collect and remotely identify host-seeking mosquitoes in field settings where the need of expert entomologists carrying out morphological identification may raise feasibility and budget concerns. To this aim, two types of traps widely used to collect host-seeking mosquitoes (i.e. the BG-Sentinel and CDC light trap) have been equipped within a sticky-cup fit for immobilizing collected specimens on a bi-dimensional surface suitable for digital image acquisition. The performance of the two traps in two villages of Burkina Faso was compared both in indoor and outdoor environments, assessing the efficiency of the identifications based on digital images with that of microscopic identification, taken as reference method.
A total of 1,519 mosquitoes belonging to 16 species was collected and 1,345 (88.5%) of these were microscopically identified as Anopheles gambiae s.l. Other species collected were mostly Culex decens (3.8%), Anopheles coustani (2.8%), and Cx. quinquefasciatus (1.6%, see Additional file 4). Ninety-eight percent of An. gambiae s.l. specimens were successfully molecularly identified as An. coluzzii (85.7%), An. arabiensis (7.1%) and An. gambiae s.s. (6.8%). Three An. coluzziiXgambiae and one An. coluzziiXarabiensis hybrids were also identified.
Mean estimates of Anopheles gambiae complex species collected by BG and CDC traps
An. gambiae s.s.
Mean estimates of Anopheles coluzzii collected by BG and CDC traps during rainy and dry seasons
These results are the first evidence from a field study showing that BG-Sentinel trap can be effectively exploited to collect host-seeking An. gambiae mosquitoes, at least when baited with BG-lure and CO2. In fact, outdoors BG trap performance was shown to be largely superior to equally baited CDC trap both in high and low mosquito density conditions, while indoors it remained effective although with lower performance compared to CDC trap. Notably, the numbers of An. coluzzii females collected per trap/night by BG trap were in the same ranges of those collected indoor by back-pack aspirations and outdoors by aspiration in pit-shelter aspirations in the same villages and seasons .
A total of 68 images with 287 mosquitoes were visually inspected and results were compared to those obtained by microscopic identification, resulting in overall success rates of 97.9% for Culicidae (281/287), 95.6% Anophelinae (262/274), 81.5% for An. gambiae s.l. (212/260) and in a success rate per single daily trap capture ranging from 75% to 100%. The images collected by capture system here described demonstrated to be detailed enough to obtain good identifications (see Additional files 1, 2, and 3). If needed, a more sensitive system could be applied, increasing the image sensor performance (i.e. more than 21Mpixels) and/or the magnification rate of lens (i.e. more than 1:1). However, the latter solution have the limit to risk to not cover the whole sticky surface, thus needing the acquisition of more than one image per sample. It is also possible to adopt the digital image identification protocol without using the glue to collect the specimens. In this case the sample collected with the classic bag can be placed on a white surface and them photographed, thus obtaining images of specimens not altered by the glue. However, the glue has the advantage to collect specimens that do not need to be killed in the field (by chloroform, ether or freezing) prior to be photographed. Moreover the image capture of glued specimens is not disturbed by the presence of wind which could move or disperse the specimens during the process. Also, the glue has the advantage to prevent possible power supply breakdowns that would accidentally allow captured insects to escape from the trap.
The type of trap used for collecting the specimens on the sticky surface (BG vs CDC) did not affect significantly species identification (Pearson’s Chi-square = 1.39, P = 0.24). This means that the position of the suction fan, which is placed after the sticky-cup in BG and before it in CDC, does not alter the quality of the collected sample. The high success rate of microscopic and digital identifications of specimens collected in sticky-cups may represent a general advantage of these modified trapping tools with respect to the unmodified traps in which the specimens are collected inside a bag. In fact, the continuous air circulation generated by the fan and the simultaneous collection of other bigger insects (e.g. moths) inside the bag may alter or even destroy diagnostic characters of mosquitoes collected, while specimens immobilized in the sticky-cup are protected from this inconvenient. The results also confirm that the glue used to immobilize the mosquitoes does not have a repellent effect, as already shown in other studies where the same glue was used on different sticky trap models to collect An. gambiae, Cx. pipiens, Ae. albopictus and Ae. aegypti in Africa, Europe and Asia [7,16,26-28].
Results strongly encourage the use of BG-trap - which is quickly becoming the new gold standard for collecting host-seeking Dengue vectors - also for collecting host-seeking An. gambiae particularly in the outdoor environment. In fact, outdoor collection of host-seeking malaria vectors have been traditionally carried out by human landing catches which are nowadays considered unethical because of the risk of exposure of human volunteers to mosquito–borne diseases. On the other hand, CDC light-trap, which is the gold standard for indoor trapping, has shown weak efficacy outdoors  and little information is presently available on the outdoor performance of more recently developed traps, such as MM-X and SUNA traps, which have been comparatively tested against other traps only in semi-field conditions [30-32]. The shown efficacy of BG-trap outdoors provides new perspectives to the challenge of collecting the outdoor biting fraction of the malaria vectors populations, whose epidemiological relevance is increasing due to the success of large-scale implementation of indoor malaria vector control strategies .
Finally, the results here obtained demonstrate that the transmission of digital images of mosquito specimens efficiently allows “remote identification” of malaria vectors, thus opening the perspective to carry out mosquito monitoring also in the absence of entomologists directly in the field. This “remote” monitoring system could be in principle further expanded to target all fractions of a vector population, exploiting already available sticky-traps for ovipositing [7,8,10,11,17], resting  and sugar-feeding  mosquitoes, as well as modifying other traps with sticky surfaces, as done in the present study.
We thank the inhabitants of Goden and Koubri for their collaboration during the whole course of this study. We also thank Paola De Santis for reviewing the English language of the manuscript. Financial support was provided by Ministero della Difesa - Ispettorato Generale della Sanità Militare, project “Telesorveglianza Vettoriale” and by the EU-FP7 grant no: 265660 “AvecNet”.
- Mallard F, Le Bourlot V, Tully T. An automated image analysis system to measure and count organisms in laboratory microcosms. PLoS One. 2013;8, e64387.View ArticlePubMed CentralPubMedGoogle Scholar
- Green JM, Appel H, Rehrig EM, Harnsomburana J, Chang J-F, Balint-Kurti P, et al. PhenoPhyte: a flexible affordable method to quantify 2D phenotypes from imagery. Plant Methods. 2012;8:45.View ArticlePubMed CentralPubMedGoogle Scholar
- Dietrich C, Hart J, Raila D, Ravaioli U, Sobh N, Sobh O, et al. InvertNet: a new paradigm for digital access to invertebrate collections. Zookeys. 2012;209:165–81.View ArticlePubMedGoogle Scholar
- Mantle BL, Salle JL, Fisher N. Whole-drawer imaging for digital management and curation of a large entomological collection. Zookeys. 2012;209:147–63.View ArticlePubMedGoogle Scholar
- Schmidt S, Balke M, Lafogler S. DScan - a high-performance digital scanning system for entomological collections. Zookeys. 2012;209:183–91.View ArticlePubMedGoogle Scholar
- Wheeler Q, Bourgoin T, Coddington J, Gostony T, Hamilton A, Larimer R, et al. Nomenclatural benchmarking: the roles of digital typification and telemicroscopy. Zookeys. 2012;209:193–202.View ArticlePubMedGoogle Scholar
- Facchinelli L, Valerio L, Pombi M, Reiter P, Costantini C, Della Torre A. Development of a novel sticky trap for container-breeding mosquitoes and evaluation of its sampling properties to monitor urban populations of Aedes albopictus. Med Vet Entomol. 2007;21:183–95.View ArticlePubMedGoogle Scholar
- Eiras AE, Buhagiar TS, Ritchie SA. Development of the gravid Aedes trap for the capture of adult female container-exploiting mosquitoes (Diptera: Culicidae). J Med Entomol. 2014;51:200–9.View ArticlePubMedGoogle Scholar
- Mackay AJ, Amador M, Barrera R. An improved autocidal gravid ovitrap for the control and surveillance of Aedes aegypti. Parasit Vectors. 2013;6:225.View ArticlePubMed CentralPubMedGoogle Scholar
- Resende M, Ázara T. Field optimisation of MosquiTRAP sampling for monitoring Aedes aegypti Linnaeus (Diptera: Culicidae). Mem Inst Oswaldo Cruz. 2012;107:294–302.View ArticlePubMedGoogle Scholar
- Ritchie SA, Long S, Hart A, Webb CE, Russell RC. An adulticidal sticky ovitrap for sampling container-breeding mosquitoes. J Am Mosq Control Assoc. 2003;19:235–42.PubMedGoogle Scholar
- Leon N. The use of sticky ovitraps to estimate dispersal of Aedes aegypti in Northeastern Mexico. J Am Mosq Control Assoc. 2001;17:93–7.Google Scholar
- Kay BH, Sutton KA, Russell BM. A sticky entry-exit trap for sampling mosquitoes in subterranean habitats. J Am Mosq Control Assoc. 2000;16:262–5.PubMedGoogle Scholar
- De Santos EMM, de Melo-Santos MAV, de Oliveira CMF, Correia JC, de Albuquerque CMR. Evaluation of a sticky trap (AedesTraP), made from disposable plastic bottles, as a monitoring tool for Aedes aegypti populations. Parasit Vectors. 2012;5:195.View ArticlePubMed CentralPubMedGoogle Scholar
- Zhang L-Y, Lei C-L. Evaluation of sticky ovitraps for the surveillance of Aedes (Stegomyia) albopictus (Skuse) and the screening of oviposition attractants from organic infusions. Ann Trop Med Parasitol. 2008;102:399–407.View ArticlePubMedGoogle Scholar
- Pombi M, Guelbeogo WM, Kreppel K, Calzetta M, Traoré A, Sanou A, et al. The sticky resting Box, a new tool for studying resting behaviour of afrotropical malaria vectors. Parasit Vectors. 2014;7:247.View ArticlePubMed CentralPubMedGoogle Scholar
- Harris C, Kihonda J, Lwetoijera D, Dongus S, Devine G, Majambere S. A simple and efficient tool for trapping gravid Anopheles at breeding sites. Parasit Vectors. 2011;4:125.View ArticlePubMed CentralPubMedGoogle Scholar
- Müller GC, Beier JC, Traore SF, Toure MB, Traore MM, Bah S, et al. Field experiments of Anopheles gambiae attraction to local fruits/seedpods and flowering plants in Mali to optimize strategies for malaria vector control in Africa using attractive toxic sugar bait methods. Malar J. 2010;9:262.View ArticlePubMed CentralPubMedGoogle Scholar
- Smallegange RC, Schmied WH, van Roey KJ, Verhulst NO, Spitzen J, Mukabana WR, et al. Sugar-fermenting yeast as an organic source of carbon dioxide to attract the malaria mosquito Anopheles gambiae. Malar J. 2010;9:292.View ArticlePubMed CentralPubMedGoogle Scholar
- Gillies M, Coetzee M. A Supplement to the Anophelinae of Africa South of the Sahara. Publ South African Inst Med Res. 1987;55:63.Google Scholar
- Santolamazza F, Mancini E, Simard F, Qi Y, Tu Z, della Torre A. Insertion polymorphisms of SINE200 retrotransposons within speciation islands of Anopheles gambiae molecular forms. Malar J. 2008;7:163.View ArticlePubMed CentralPubMedGoogle Scholar
- Crawley M. The R Book. London: John Wiley & Sons, Inc; 2013. p. 1051.Google Scholar
- R Core Team. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2013.Google Scholar
- Bates D, Maechler M, Bolker B. Linear mixed-effects models using S4 classes. 2013. Maintainer: lme4-author@R-forge.wu-wien.ac; http://lme4.r-forge.r-project.org.
- Skaug H, Fournier D, Bolker B, Magnusson A, Nielsen A. Generalized Linear Mixed Models using AD Model Builder. R package. 2014. Maintainer: firstname.lastname@example.org; http://glmmadmb.r-forge.r-project.org/repos; http://admb-project.org.
- Facchinelli L, Koenraadt CJM, Fanello C, Kijchalao U, Valerio L, Jones JW, et al. Evaluation of a sticky trap for collecting Aedes (Stegomyia) adults in a Dengue-endemic area in Thailand. Am J Trop Med Hyg. 2008;78:904–9.PubMedGoogle Scholar
- Marini F, Caputo B, Pombi M, Tarsitani G, della Torre A. Study of Aedes albopictus dispersal in Rome, Italy, using sticky traps in mark-release-recapture experiments. Med Vet Entomol. 2010;24:361–8.View ArticlePubMedGoogle Scholar
- Valerio L, Marini F, Bongiorno G, Facchinelli L, Pombi M, Caputo B, et al. Host-feeding patterns of Aedes albopictus (Diptera: Culicidae) in urban and rural contexts within Rome province, Italy. Vector Borne Zoonotic Dis. 2010;10:291–4.View ArticlePubMedGoogle Scholar
- Silver JB. Mosquito Ecology: Field Sampling Methods. Thirdth ed. London: Springer; 2008. p. 1498.View ArticleGoogle Scholar
- Schmied WH, Takken W, Killeen GF, Knols BGJ, Smallegange RC. Evaluation of two counterflow traps for testing behaviour-mediating compounds for the malaria vector Anopheles gambiae s.s. Under semi-field conditions in Tanzania. Malar J. 2008;7:230.View ArticlePubMed CentralPubMedGoogle Scholar
- Hiscox A, Otieno B, Kibet A, Mweresa CK, Omusula P, Geier M, et al. Development and optimization of the Suna trap as a tool for mosquito monitoring and control. Malar J. 2014;13:257.View ArticlePubMed CentralPubMedGoogle Scholar
- Nyasembe VO, Tchouassi DP, Kirwa HK, Foster WA, Teal PEA, Borgemeister C, et al. Development and assessment of plant-based synthetic odor baits for surveillance and control of malaria vectors. PLoS One. 2014;9, e89818.View ArticlePubMed CentralPubMedGoogle Scholar
- World Health Organization. World Malaria Report 2013. Geneva, Switzerland: World Health Organization; 2013. p. 284.Google Scholar
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.