Orsborne et al. Parasites Vectors (2019) 12:143 https://doi.org/10.1186/s13071-019-3401-3 Parasites & Vectors RESEARCH Open Access Investigating the blood-host plasticity and dispersal of Anopheles coluzzii using a novel field-based methodology James Orsborne1, Luis Furuya‑Kanamori2,3, Claire L. Jeffries1, Mojca Kristan1, Abdul Rahim Mohammed4, Yaw A. Afrane4, Kathleen O’Reilly1, Eduardo Massad5, Chris Drakeley6, Thomas Walker1 and Laith Yakob1* Abstract Background: The biting behaviour and dispersal of insect vectors in the field underlies the transmission of many diseases. Here, a novel collection methodology coupled with the molecular analysis of blood‑meal sources and diges‑ tion rates is introduced with the aim of aiding the understanding of two critical and relatively understudied mosquito behaviours: plasticity in blood‑host choice and vector dispersal. Results: A collection strategy utilising a transect of mosquito traps placed at 50 m intervals allowed the collection of blood‑fed Anopheles coluzzii from a malaria‑endemic village of southern Ghana where human host availability ranged from zero (a cattle pen), increasing until humans were the dominant host choice (the middle of the village). Blood‑meal analysis using PCR showed statistically significant variation in blood‑meal origins for mosquitoes collected across the 250 m transect: with decreasing trend in Bovine Blood Index (OR = 0.60 95% CI: 0.49–0.73, P < 0.01) and correspondingly, an increasing trend in Human Blood Index (OR = 1.50 95% CI: 1.05–2.16, P = 0.028) as the transect approached the village. Using qPCR, the host DNA remaining in the blood meal was quantified for field‑caught mosquitoes and calibrated according to timed blood digestion in colony mosquitoes. Time since blood meal was consumed and the corresponding distance the vector was caught from its blood‑host allowed the estimation of An. coluzzii dispersal rates. Within 7 hours of feeding, mosquitoes typically remained within 50 m of their blood‑host but at 60 hours they had dispersed up to 250 m. Conclusions: Using this methodology the remarkably small spatial scale at which An. coluzzii blood‑host choice can change was demonstrated. In addition, conducting qPCR on host blood from field‑caught mosquitoes and calibrat‑ ing with timed experiments with colonised mosquitoes presents a novel methodology for investigating the dispersal behaviour of vectors. Future adaptations to this novel method to make it broadly applicable to other types of setting are also discussed. Keywords: Blood‑meal analysis, Host preference, Mosquito, Biting preference, Blood index Background seek a proportion of blood meals from alternative (non- Many disease vectors have demonstrable preference human) host sources [1–4]. Gillies first researched host for a particular type of mammalian host to obtain a choice among malaria vectors by releasing Anopheles blood meal, however it is well documented that even mosquitoes into an enclosed space and comparing the the most anthropophilic of disease vectors will still numbers flying into a room holding a human volunteer with those entering a room with a calf [5]. In the subse- quent 50  years, the complexity of host preference and *Correspondence: Laith.yakob@lshtm.ac.uk biting behaviour has become well documented [2, 6, 7]. 1 Department of Disease Control, London School of Hygiene & Tropical While useful, many host-choice experiments have set- Medicine, London, UK Full list of author information is available at the end of the article ups that can only inform the intrinsic host preference © The Author(s) 2019. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creat iveco mmons .org/ publi cdomai n/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Orsborne et al. Parasites Vectors (2019) 12:143 Page 2 of 8 of a vector, and may or may not be indicative of what for haematophagous disease vectors. Finally, potential host species is bitten in natural field settings [8]. future adaptations to these novel methods are discussed Many extrinsic as well as intrinsic factors play a part in order to make them broadly applicable to investigating in who or what is ultimately bitten by a disease vec- host plasticity and dispersal in other settings. tor in a field setting, and these have been summarised comprehensively [2]. This balance between intrinsic Methods and extrinsic factors could go some way to explaining Study site and mosquito collection the large variability found in the reported human blood Mosquitoes were collected from the village of Dogo, index (HBI) of major disease vectors [2, 9]. Although it in the Greater Accra region of Ghana (05°52.418N, has been recognised for a long time that the same mos- 00°33.607E). The village is in the south-eastern coast quito population will often adjust its biting towards a of Ghana, with the Gulf of Guinea to the south and the more locally available host species [7, 9, 10], the extent Volta River to the east. The average rainfall is approxi- of this plasticity and the spatial scale at which it acts mately 927 mm per year with the main rainy season from remains understudied even for the most important dis- April to June and a shorter second season in October. ease vectors. This plasticity is an important factor when Temperatures range between 23–33  °C. The area is cos- it comes to implementing control strategies. The intro- tal savannah with sandy soil, short savannah grass with duction of insecticide treated nets (ITNs) and indoor some small/medium sized trees. The land is used exten- residual spraying (IRS) has seen the biting behaviour sively for grazing livestock as well as growing crops for of many major malaria vectors shift [2, 11, 12] with local trade. Housing mostly consisted of concrete struc- increasing reports of these vectors seeking blood-meals tures with concrete/brick walls and flooring. Some tradi- from alternative non-human sources [3]. Outdoor and tional mud style houses were also present, more so on the residual malaria transmission supported by secondary periphery of the village. or indiscriminate malaria vectors [13] further high- Mosquitoes were collected across five consecutive lights the importance of understanding host choice so nights in June 2017. The trapping setup consisted of CDC future control strategies can be better targeted. resting traps placed outdoors at 50 m intervals form- Also implicit to the spatial scale across which feeding ing a 250  m transect comprising of six trapping points choice changes is the vector’s dispersal ability. For exam- (denoted T1–T6). This transect was set beginning at an ple, if a vector tends not to disperse very far, a reason- area of zero human population density (T1, outside of the able assumption may be that it will be less discerning village by cattle resting and overnight holding pens) and in its choice of host and therefore be more likely to bite extending towards a human population in 50 m intervals whatever is nearby. However, what is of considerable ending at an area of high human density (see Table 1 for hindrance to this field’s development is the absence of description and Fig. 1 for map of collection site). Mosqui- reliable methods for assessing a disease vector’s disper- toes were collected overnight from 18:00 to 06:00 h. sal ability. Conducting experimental studies on mos- Mosquitoes were removed from traps at 06:00 h each quito dispersal has been particularly challenging with morning and immediately killed using chloroform to stop the majority of such experiments involving the mark- any active blood-meal digestion. Mosquitoes where then release-recapture of mosquitoes. However, the impact sorted with all blood-fed females being processed first. of handling mosquitoes combined with the typically low All visually blood-fed and gravid Anopheles mosquitoes recapture rates (in the order of < 2% for An. gambiae were processed individually with transect location and [14–18]) has limited what can be learned. night collected being recorded. Abdomens of blood- Here, the blood-meal sources were identified for An. fed mosquitoes were removed with sterile forceps and coluzzii caught in traps situated across a 250  m tran- pressed onto FTA®Classic cards (Whatman, GE Health- sect representing a range of alternative blood-host spe- care, New Jersey, USA) to preserve the blood meal for cies availabilities (primarily human or cattle) in a malaria molecular analysis. Excess blood-fed mosquitoes were endemic village of southern Ghana. By using this collec- preserved in RNA later (Thermo Fisher Scientific Life tion methodology coupled with molecular blood-meal Technologies, Massachusetts, USA) in a 96-well plate identification, we aim to investigate the spatial range where necessary. across which this principle malaria vector can adjust its targeted blood-host species based on local host avail- DNA extraction ability. In addition, by quantifying host DNA isolated Mosquito abdomens were extracted individually. Samples from field-caught vectors and calibrating this with timed were homogenised using a Qiagen TissueLyser II (Qia- laboratory mosquito feeding experiments, an alterna- gen, Manchester, UK) with a 5 mm stainless steel bead tive method is presented for measuring dispersal rates (Qiagen) placed in each sample tube in a 96-well plate Orsborne et al. Parasites Vectors (2019) 12:143 Page 3 of 8 Table 1 Description of areas around transects where mosquitoes were collected including number and type of host present Transect Description Approximate no. of hosts 1 Evening holding pen for cattle for the village (used from 18:00 h to 06:00 h), one small uninhab‑ Cows (n = 150) ited house next to the pen was used to hold tools and supplies for cattle farmers 2 End of cattle pen (as described above), small pig holding and empty cattle shed. Edge of village Cows (n = 150); pigs (n = 5) is approximately 30 m away with empty newly built houses; first house with inhabitants (T3) 50 m away 3 First cluster of 4 small households on periphery of village c.50 m from cattle pen. A small hold‑ Humans (n = approx. > 20); chickens ing of chickens and goats as well as pet dogs which roam the area freely (n = 7); dogs (n = 3); goats (n = 4) 4 Complex of 5 houses, 3 guinea fowl and 2 cats present, guinea fowl nested in nearby outbuild‑ Humans (n = approx. > 30); guinea fowl ing, cats roamed freely (n = 3); cats (n = 2) 5 Complex of 8 houses, no fixed animal housing Humans (n = approx. > 45) 6 Dogo village, and the largest density of households; one small chicken coop but no other Humans (n = approx. > 85); chickens (n = 3) animal holdings, no dogs or cats seen Fig. 1 Map of collection site and host species present at each transect point (transect 250 m in total) taken from Google Earth Pro format. Once homogenised, DNA was then extracted punch. Resulting punches were incubated in ATL buffer using the Qiagen DNeasy 96 kits (Qiagen) following and Proteinase K for 6 h before DNA extraction was manufacturer’s protocol. Blood meals preserved on FTA performed following manufacturer’s protocol. Extracted cards were punched out using a sterile steel 4 mm radius DNA was stored at -20 °C until analysed. Orsborne et al. Parasites Vectors (2019) 12:143 Page 4 of 8 Mosquito species identification available host species in the area. The reaction conditions Mosquito species identification was initiated using a consisted of a 10 µl reaction including 0.5 M of forward real-time multiplex PCR assay targeting the rRNA gene and reverse primers (Integrated DNA Technologies, Leu- [12]. Standard forward and reverse primers were used in ven, Belgium), 5  µl of SYBR green master mix (Roche, conjunction with two species-specific Taqman probes. Welwyn Garden City, UK), 2  µl of nuclease-free water The reaction conditions were as follows: a 12.5  µl reac- (Roche) and 2  µl of template DNA. PCR was run on a tion containing 1 µl of genomic DNA. 6.25 µl of Quan- LightCycler 96 real-time PCR machine (Roche) under the tinova (Qiagen) probe master mix, 800  nM of forward following cycling conditions: pre-incubation of 95 °C for and reverse primers (Thermo Fischer Scientific, East 600 s, 40 cycles of 95 °C for 10 s, 62 °C for 10 s and 72 °C Grinstead, UK), 200 nM of An. arabiensis probe (Sigma- for 30 s followed by a melting analysis. Aldrich, Gillingham, UK) and 80  nM of An. gambiae Human-positive blood meals (including any potentially probe (Applied Biosystems, UK). Samples were run on a mixed feeds) from the above assay were confirmed using Stratagene MX3005P (Agilent Technologies, Santa Clara, the Promega Plexor® HY Human DNA forensic detection USA) using cycling conditions of 10  min at 95  °C, fol- kit (Promega, Southampton, UK). Assay was performed lowed by 40 cycles of 95  °C for 25 s and 66  °C for 60 s. following manufacturer’s protocol using a Stratagene The increases in fluorescence were monitored in real MX3005P (Agilent Technologies, Santa Clara, USA) real- time by acquiring at the end of each cycle. time PCR machine. To differentiate between An. coluzzii and An. gambiae within the An. gambiae species complex, a single end- Laboratory assessment of blood‑meal DNA degradation point PCR was performed. This PCR targets the SINE200 rate retrotransposon and utilising an insertion in this area Approximately 500 female An. coluzzii mosquitoes allows the two species to be distinguished following gel (N’gousso strain originally collected from Yaounde, Cam- visualisation [13]. Anopheles coluzzii produces a band eroon) were placed into an insect cage (Bugdorm, Wat- at 479  bp with An. gambiae producing a band at 249 kins and Doncaster, UK) and, using a Hemotek, fed for bp. Reaction was as follows: a 25 µl reaction containing 15 min on bovine blood collected from a UK based abat- 0.5 mM of forward (5′-TCG CCT TAG ACC TTG CGT toir (First Line UK (Ltd), UK). Mosquitoes were reared TA-3′) and reverse (5′-CGC TTC AAG AAT TCG AGA at the London School of Hygiene & Tropical Medicine TAC-3′) primers, 12.5  µl of Hot start Taq polymerase under standardized conditions in an incubator (27 °C and (New England Biolabs, Ipswich, UK), 9.5 µl of nuclease- 70% humidity with a 12:12 light/dark photocycle) and free water and 2 µl of template DNA. Cycling conditions given access to 10% sugar solution. Female mosquitoes were as follows: 10 min at 94 °C followed by 35 cycles of were individually collected and checked for feeding sta- 94 °C for 30 s, 54 °C for 30 s, 72 °C for 60 s, and a final tus. Only overtly fully fed mosquitoes were selected for elongation step of 72 °C for 10 min. the experiment. Fully-fed females were separated into PCR products were visualised on a 2% agarose gel using paper cups covered with netting; each cup contained an Egel E-Gel iBase Power System and E-Gel Safe Imager a maximum of 30 female mosquitoes. Every 6 hours a Real-Time Transilluminator (Invitrogen, East Grin- single cup was removed and placed in a -80 °C freezer stead, UK). The assay was performed on 10% of all sam- to kill the mosquitoes and stop blood-meal digestion. ples identified as An. gambiae from the first assay with This was repeated until the mosquitoes had completely corresponding controls. Samples producing unknown digested the blood-meal or were visually gravid. DNA or inconclusive results were sequenced (ITS2 Sanger was extracted using the above protocol from seven whole sequencing) using primers originally developed by Beebe bodies for each time point. A 1:10 serial dilution of all & Saul [19] and sequences were used to perform nucle- time = 0 samples was used to generate a standard curve otide BLAST (NCBI) database queries. PCR reactions with dilutions being made down to 1 × 10-7. The standard were performed on a T100 Thermal Cycler (Bio-Rad Lab- curve was used to assess assay sensitivity (limit of detec- oratories, Watford, UK) and amplified gene fragments tion) with the resulting Ct values from each time point were visualized by electrophoresis on a 2% agarose gel being used to estimate the time post-blood-feed for the using an E-gel E-Gel iBase Power System and E-Gel Safe field-caught mosquitoes. DNA from the blood meals Imager Real-Time Transilluminator (Invitrogen). from the field-caught mosquitoes was quantified using the same protocol. As larger female mosquitoes typically Blood‑meal identification obtain a larger blood meal when feeding [21], we nor- Samples were initially screened using bovine and human malised for mosquito body size to account for the pos- specific primers developed by Gunathilaka et  al. [20]. sibility that the different quantity of bovine DNA across These primers were selected based on the abundance of the transect was due to mosquito size rather than time Orsborne et al. Parasites Vectors (2019) 12:143 Page 5 of 8 post blood-meal. Ct values for bovine DNA were nor- Sanger sequencing as An. melas and was excluded from malised against the Ct values for the corresponding host the analysis (Table 2). mosquito ribosomal DNA (rDNA) gene used for spe- The dominant mosquito blood meal was of bovine cies identification, producing a ratio of bovine (Bos tau- origin with 73.5% of all meals being sourced from these rus mtDNA)-to-vector DNA (An. coluzzii rDNA). In hosts. Four (1.3%) individual mosquitoes were found to this way, the quantity of bovine DNA measured for the have solely fed on humans with an additional ten (3.3%) timed experiments with colonised mosquitoes was used having a mixed feed of both bovine and human blood to estimate the time since last blood meal of the mosqui- (Table  3). Figure  2 shows how the bovine blood index toes caught at the different transect points. In conjunc- (BBI) varied significantly across the transect, indicat- tion with the known distances between the hosts and the ing a decreasing trend with increasing distance from the transect points, this estimated time since last blood meal cattle shed (OR = 0.60, 95% CI: 0.49–0.73, P < 0.01). The informed the dispersal rate of the vectors. opposite trend was observed for human blood meals with the HBI increasing significantly towards the village Statistical analysis (OR = 1.50, 95% CI: 1.05–2.16, P = 0.028). All statistical analysis was performed using STATA and Focusing on mosquitoes that had fed on cattle PRISM. Trends in blood indices across the transect were (n = 227), it was observed that the quantity of blood- tested for the field-caught mosquitoes using a general- host DNA extracted from mosquitoes varied across the ised linear model (glm) with a binomial function. Odds transect with the average PCR cycle threshold (Ct) val- ratios were calculated for proportion of bovine or human ues for bovine blood detection being 20.72 (95% CI: fed mosquitoes across each collection night as a total of 18.98–22.45) for mosquitoes caught by the cattle pen and An. coluzzii collected and P-values (P < 0.05) were used 30.15 (23.14–37.16) for mosquitoes caught 250 m away to interpret any significant trends. Linear regression was (P < 0.01). As detection of rDNA is a proxy for total mos- performed to investigate the correlation between bovine quito DNA extracted and therefore body size, we com- Ct value and time post-feed recorded in the experiments pared the ratios of bovine-to-vector DNA (An. coluzzii with colony insects. rDNA) across the transect to ensure different quantities of bovine DNA detected at different distances from the Results hosts was not due to mosquito size but rather time post- A total of 318 blood-fed Anopheles mosquitoes were col- blood-meal. The correlation between Ct ratio and dis- lected over a five-night period. Of these, 307 were identi- tance from cattle was retained (t(225) = -2.18, P = 0.03). fied as part of the An. gambiae species complex: 306 were The experimental time series was performed with a identified as An. coluzzii using a combination of species- laboratory colony of An. coluzzii and producing mean specific PCRs and Sanger sequencing of a fragment of the Ct values for known time points post-blood-feeding. ITS2 region. The remaining insect was identified by ITS2 The time series showed Ct values increased with time Table 2 Total number of blood‑fed mosquitoes caught by species and transect point T1 T2 T3 T4 T5 T6 Total An. coluzzii 17 153 72 26 22 16 306 An. melas 0 1 0 0 0 0 1 Other species 0 9 2 0 0 0 11 Total 17 163 74 26 22 16 318 Table 3 Total number of An. coluzzii mosquitoes collected by blood‑meal source and transect point Host source T1 T2 T3 T4 T5 T6 Total % Bovine fed 15 138 39 12 18 5 227 74.18 Confirmed human feds 0 1 0 1 1 1 4 1.31 Mixed human/bovine 0 5 2 1 0 2 10 3.27 Unknown 2 9 31 12 3 8 65 21.24 Total caught 17 153 72 26 22 16 306 100 Orsborne et al. Parasites Vectors (2019) 12:143 Page 6 of 8 Discussion Evidence for the influence of local host availability on blood-host selection was demonstrated through analy- sis of the blood meals of An. coluzzii caught from the field using a novel sampling strategy. Previous investiga- tions of HBI in field settings have demonstrated its large variability across and within species [2]. The aim of the present study was to investigate, for the same mosquito population, what level of variability can be expected, and, to determine the spatial scale that this choice can vary. Here, a relatively low-cost (the chief expense being the PCR for blood-host species identification) and sim- ple experimental setup for investigating host choice Fig. 2 The human blood index (triangles) and bovine blood index in the field is described. It was demonstrated that local (circles) for each transect point (T1‑T6), along with 95% confidence host availability plays a crucial role in the host choice of intervals, for all blood‑fed An. coluzzii mosquitoes collected a major malaria vector. Moreover, the remarkably small spatial scale (~250 m) at which this behaviour can be sig- nificantly impacted is demonstrated for the first time. post-feed (P < 0.01, see Fig.  3) with no bovine DNA Results could have significant implications for vector detected after the 60-hour time point. Regression anal- control. For example, field studies involving endecto- ysis showed a positive correlation between bovine Ct cidal applications on livestock have shown encouraging value and time post-feed in the experimental time series 2 results in terms of long-lasting mosquitocidal effects [22, (R = 0.92, slope = 0.183; see Fig.  3). Calibrating the 23]. However, previously this strategy has only been con- blood-meals of field-caught mosquitoes using the timed sidered for targeting malaria vector species traditionally experiment with our mosquito colony, the dispersal rate viewed as zoophilic (e.g. An. arabiensis). Recently, this of An. coluzzii could then be extrapolated: within 7 hours assumption has been challenged by the demonstration of feeding, mosquitoes typically remained within 50  m that even the most anthropophagic populations of vec- of their blood-host but at 60 hours had dispersed up to tors readily bite non-human hosts, and that the methods 250 m (Fig. 3). for assessing host choice exclusively from mosquitoes Fig. 3 Effect of time post blood‑meal on mean bovine Ct values produced from qPCR. Shown are the means (bars indicate 95% CIs) of experimental time series (black), the serial dilution Ct values to assess assay sensitivity (blue), the mean (and 95% CI) Ct values of each transect point (red) and regression line used to predict time post feeding (dashed black line). Note that ‘time post‑feed’ is from direct observation for colony mosquito blood‑meal digestions (black) but is then extrapolated to the estimated time post‑feed for field‑caught mosquitoes Orsborne et al. Parasites Vectors (2019) 12:143 Page 7 of 8 caught in human habitation may suffer from systematic blood-fed mosquitoes caught nearby is one such setup bias [9]. Therefore, one way by which the current study that requires future investigation. adds to the discussion of optimal vector control strategy Thirdly, blood-meal digestion levels of field-caught is through the provision of a simple method for assess- mosquitoes were calibrated with colonised mosquitoes. ing the degree to which anthropophagy varies for a given Here multiple differences can occur: colony fed mosqui- mosquito population. This can then be used to inform toes are reared at controlled densities, temperature and strategies for improved targeting of different control humidity and are able to take a full blood meal without methods such as endectocides. Coupled entomological- encountering any defensive behaviour from hosts. These epidemiological modelling frameworks already exist for are of stark difference to what blood-fed mosquitoes may using these data to inform projections of this novel vec- encounter in the field. A realistic temperature/humid- tor control [24], including its use as part of an integrated ity regimen that better emulates natural diurnal patterns vector management programme [25]. has been shown to significantly impact various aspects Linking the quantity of host-blood DNA isolated from of mosquito metabolism [23]; and, artificially controlling mosquitoes caught at known distances from the specific larval density can produce mosquitoes of similar size and host species with timed blood-meal digestion assays con- fitness, something which may not be comparable to the ducted on colonised mosquitoes presents a novel method field. Future experiments to ascertain the influence that for informing dispersal rates of mosquito populations. these factors may have on blood-meal digestion would Dispersal is recognised to underlie mosquito population constitute an important next step. structure [17] as well as human exposure to transmis- sion [26] and our ability to control transmission [27]. Yet, knowledge of this critical aspect of behaviour has been Conclusions hampered by our inability to produce reliable estimates Results presented in this study provide new insight into of vector dispersal in the field. To our knowledge, this fundamental aspects of malaria vectors with important study provides the first estimates using a non-intrusive implications for malaria control strategy. Additionally, and easily repeatable method for measuring malaria vec- the novel experimental design presented offers a new tor dispersal that informs the mosquito’s dispersal rate methodology in measuring dispersal that with further across its gonotrophic cycle (approximately 2.5  days). development could be broadly applicable to other field- However, it is important to address some of the present caught blood-feeding disease vectors. study’s limitations and to identify some areas of future development of this approach. First, in this study the numbers of mosquitoes captured AbbreviationsHBI: human blood index; BBI: bovine blood index; Ct: cycle threshold; OR: odds nearby humans was low compared to those caught adja- ratio; CDC: centers for disease control and prevention. cent to cattle. That said, only 5 nights of mosquito cap- tures were needed in order for a statistically significant AcknowledgementsWe would like to thank the community of Dogo as a whole for allowing this trend to be identified for host choice across the transect. study to be performed with particular mention to the individuals who allowed In the future, increasing the duration of the experiment us to place traps around their homes. We would also like to thank Issac Adjaot‑ would improve its ability to inform the likely shape of dis- tor for allowing us access to the community and making this work possible. persal (e.g. leptokurtic versus Gaussian), something that Funding could not be achieved with the present study. JO has an MRC London Intercollegiate Doctoral Training Partnership Student‑ Secondly, in order to estimate distances from blood- ship. TW and CLJ are funded through a Wellcome Trust/Royal Society Sir Henry Dale Fellowship (101285/Z/13/Z) awarded to TW. LY received funds from a hosts these hosts must remain spatially confined. While Royal Society Research Project (RSG\R1\180203). YAA has funding from the this was possible in the present study because cattle were National Institute of Health, US (R01AI123074). LFK is supported by an Austral‑ confined to their holding pen, this may require experi- ian National Health and Medical Research Council Fellowship (APP1158469). Funding bodies had no role in the design of the study and collection, analysis mental adaptations for other types of environment. It and interpretation of data nor in writing the manuscript. must be made clear that this experiment in this particular field site was not intended to inform An. coluzzii disper- Availability of data and materialsAll data generated or analysed during this study are included in this published sal rates everywhere that this vector can be found. Rather, article. the aim of this study was to present a new method for measuring dispersal that can be adapted to other settings Authors’ contributionsLY and JO conceived the study. JO, ARM and YAA performed the field work. to inform local mosquito behaviour. For example, tether- JO, CLJ, MK and TW performed laboratory analyses. All authors contributed ing an animal species not otherwise found in the vicin- to experimental designs, results interpretation and manuscript drafting. All ity of a field site, followed by identifying its DNA from authors read and approved the final manuscript. Orsborne et al. Parasites Vectors (2019) 12:143 Page 8 of 8 Ethics approval and consent to participate 11. Ototo EN, Mbugi JP, Wanjala CL, Zhou G, Githeko AK, Yan G. Surveillance of Study protocol and all relevant documentation was reviewed by the LSHTM malaria vector population density and biting behaviour in western Kenya. Ethics Committee before the study commenced. In country ethical approval Malar J. 2015;14:244. was also obtained by YAA from the University of Ghana’s Ethics committee. 12. Thomsen EK, Koimbu G, Pulford J, Jamea‑Maiasa S, Ura Y, Keven JB, et al. Mosquito behavior change after distribution of bednets results in decreased Consent for publication protection against malaria exposure. J Infect Dis. 2017;215:790–7. Not applicable. 13. Killeen GF, Kiware SS, Okumu FO, Sinka ME, Moyes CL, Massey NC, et al. Going beyond personal protection against mosquito bites to eliminate Competing interests malaria transmission: population suppression of malaria vectors that exploit The authors declare that they have no competing interests. both human and animal blood. BMJ Glob Health. 2017;2:e000198. 14. Costantini C, Li S‑G, Torre AD, Sagnon NF, Coluzzi M, Taylor CE. Density, Publisher’s Note survival and dispersal of Anopheles gambiae complex mosquitoes in a West African Sudan savanna village. Med Vet Entomol. 1996;10:203–19. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ 15. Takken W, Charlwood JD, Billingsley PF, Gort G. Dispersal and survival of lished maps and institutional affiliations. Anopheles funestus and A. gambiae s.l. (Diptera: Culicidae) during the rainy season in southeast Tanzania. Bull Entomol Res. 2009;88:561–6. Author details 16. Midega JT, Mbogo CM, Mwambi H, Wilson MD, Ojwang G, Mwangangi JM, 1 Department of Disease Control, London School of Hygiene & Tropical Medi‑ et al. Estimating dispersal and survival of Anopheles gambiae and Anopheles cine, London, UK. 2 Department of Population Medicine, College of Medicine, funestus along the Kenyan coast by using mark‑release‑recapture methods. Qatar University, Doha, Qatar. 3 Research School of Population Health, College J Med Entomol. 2007;44:923–9. of Health and Medicine, Australian National University, Canberra, Australia. 17. Thomson MC, Connor SJ, QuiÑOnes ML, Jawara M, Todd J, Greenwood BM. 4 Department of Medical Microbiology, College of Health Sciences, University Movement of Anopheles gambiae s.l. malaria vectors between villages in The of Ghana, Korle Bu, Accra, Ghana. 5 School of Applied Mathematics, Fundacao Gambia. Med Vet Entomol. 1995;9:413–9. Getulio Vargas, Rio de Janeiro, Brazil. 6 Department of Immunology & Infection, 18. Gillies MT. Studies on the dispersion and survival of Anopheles gambiae Giles London School of Hygiene & Tropical Medicine, London, UK. in East Africa, by means of marking and release experiments. Bull Entomol Res. 1961;52:99–127. Received: 5 October 2018 Accepted: 15 March 2019 19. Beebe NW, Saul A. Discrimination of all members of the Anopheles punctu- latus complex by polymerase chain reaction‑restriction fragment length polymorphism analysis. Am J Trop Med Hyg. 1995;53:478–81. 20. Gunathilaka N, Denipitiya T, Hapugoda M, Abeyewickreme W, Wickremas‑ inghe R. Determination of the foraging behaviour and blood meal source References of malaria vector mosquitoes in Trincomalee District of Sri Lanka using a 1. White GB, Magayuka SA, Boreham PFL. Comparative studies on sibling spe‑ multiplex real time polymerase chain reaction assay. Malar J. 2016;15:242. cies of the Anopheles gambiae Giles complex (Dipt., Culicidae): bionomics 21. Takken W, Klowden MJ, Chambers GM. Effect of body size on host seeking and vectorial activity of species A and species B at Segera, Tanzania. Bull and blood meal utilization in Anopheles gambiae sensu stricto (Diptera: Entomol Res. 1972;62:295–317. Culicidae): the disadvantage of being small. J Med Entomol. 1998;35:639–45. 2. Takken W, Verhulst NO. Host preferences of blood‑feeding mosquitoes. 22. Chaccour CJ, Ngha’bi K, Abizanda G, Irigoyen Barrio A, Aldaz A, Okumu F, Annu Rev Entomol. 2013;58:433–53. et al. Targeting cattle for malaria elimination: marked reduction of Anopheles 3. Sousa CA, Pinto J, Almeida PG, Ferreira C, Do Rosário VE, Charlwood JD. arabiensis survival for over six months using a slow‑release ivermectin Dogs as a favored host choice of Anopheles gambiae sensu stricto (Diptera: implant formulation. Parasit Vectors. 2018;11:287. Culicidae) of São Tomé, West Africa. J Med Entomol. 2001;38:122–5. 23. Poche R, Burruss D, Polyakova L, Poche D, Garlapati R. Treatment of livestock 4. Dumonteil E, Ramirez‑Sierra M‑J, Pérez‑Carrillo S, Teh‑Poot C, Herrera C, with systemic insecticides for control of Anopheles arabiensis in western Gourbière S, et al. Detailed ecological associations of triatomines revealed Kenya. Malar J. 2015;14:351. by metabarcoding and next‑generation sequencing: implications for 24. Yakob L. Endectocide‑treated cattle for malaria control: a coupled entomo‑ triatomine behavior and Trypanosoma cruzi transmission cycles. Sci Rep. logical‑epidemiological model. Parasite Epidemiol Control. 2016;1:2–9. 2018;8:4140. 25. Yakob L, Cameron M, Lines J. Combining indoor and outdoor methods for 5. Gillies MT. Selection for host preference in Anopheles gambiae. Nature. controlling malaria vectors: an ecological model of endectocide‑treated 1964;203:852. livestock and insecticidal bed nets. Malar J. 2017;16:114. 6. Lyimo IN, Haydon DT, Russell TL, Mbina KF, Daraja AA, Mbehela EM, et al. The 26. Service MW. Mosquito (Diptera: Culicidae) dispersal ‑ the long and short of impact of host species and vector control measures on the fitness of African it. J Med Entomol. 1997;34:579–88. malaria vectors. Proc R Soc B. 2013;280:20122823. 27. Carter R, Mendis KN, Roberts D. Spatial targeting of interventions against 7. Garrett‑Jones C. The human blood index of malaria vectors in relation to malaria. Bull World Health Organ. 2000;78:1401–11. epidemiological assessment. Bull World Health Organ. 1964;30:241–61. 8. Besansky NJ, Hill CA, Costantini C. No accounting for taste: host preference in malaria vectors. Trends Parasitol. 2004;20:249–51. 9. Orsborne J, Furuya‑Kanamori L, Jeffries CL, Kristan M, Mohammed AR, Afrane YA, et al. Using the human blood index to investigate host biting plasticity: a systematic review and meta‑regression of the three major Ready to submit your research ? Choose BMC and benefit from: African malaria vectors. Malar J. 2018;17:479. 10. Chandler JA, Boreham PF, Highton RB, Hill MN. A study of the host selection • fast, convenient online submission patterns of the mosquitoes of the Kisumu area of Kenya. Trans R Soc Trop • thorough peer review by experienced rese archers in your field Med Hyg. 1975;69:415–25. • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions