Introduction

Vector control remains a cornerstone of malaria prevention, with vector surveillance playing a critical role in monitoring and evaluating control and research programs. However, for surveillance to be truly informative, high-quality data must be collected consistently across space and time, extending beyond simple vector densities and morphological identification. A major challenge is that molecular assays (essential for understanding vector species composition) are prohibitively expensive when applied to thousands of mosquitoes annually. As countries progress toward malaria elimination, gaining deeper insights into vectorial systems and transmission patterns becomes essential for designing end-game strategies. This underscores the urgent need for cost-effective tools that enable large-scale sample processing, ensuring robust data collection to drive informed decision-making in malaria control and elimination efforts.

MALDI-TOF MS represents one such tool. The technology has revolutionised diagnostics in the field of clinical microbiology allowing for faster turn-around time from sample preparation to obtaining precise and accurate results at a lower cost per sample [1–4]. Recently, MALDI-TOF has found valuable application in entomological surveillance. Numerous studies have assessed its performance, accuracy, and reliability for mosquito species and sub-species identification [5–8]. MALDI-TOF MS has also been evaluated for the identification of other arthropods, including Drosophila, [9,10], fleas [11], lice [12], ticks [13–16], Ceratopogonides (biting midges), and sand flies [17–19]. It has also has also been used to differentiate between infected and non-infected ticks [15,16,20]; Plasmodium-infected and non-infected An. stephensi [21] and Wolbachia-infected and non-infected Aedes aegypti [22]. MALDI-TOF MS has also been shown to distinguish between/among blood meals from different hosts [5,23–25], age grading of malaria vectors [26] and insecticide resistance profiling [27].

Creating in house libraries can be challenging and often the number of institutions that have access to the instrument but that use it beyond microbiology is limited. We aim to demystify the process and share a detailed workflow for creating a MALDI-TOF MS library using malaria vector species identification as a case study. When well optimised, the spectra obtained from a single run can be used to develop other libraries and query other parameters including, for example, infectious status [21], age/parity [26], presence endosymbionts such as Microsporidia MB or Wolbachia [22], and potentially insecticide resistance [27]. Blood-engorged mosquito abdomens can be used for differentiation of blood meal sources [5].

The workflow involves a series of steps that begins with dissection of mosquitoes into three anatomical compartments: legs and wings, cephalothorax, and abdomen. Dissection is followed by protein extraction and acquisition of spectral profiles. Guided by results from gold standard tests (ELISA, PCR, taxonomy), a sub-set of high-quality spectra is then selected and used for training the machine algorithm and create a reference database (spectral library) (training dataset).

A high-quality spectrum should have a high number of well-defined, intense peaks with good signal-to-noise ratio, consistent peak positions across replicates, minimal background noise, and a set of prominent “marker peaks” specific to the analysed molecule, allowing for reliable identification and comparison against a reference database. The remaining spectra are used to determine the accuracy of the reference database in identifying unknown samples (validation dataset). The match level between the unknown sample and the reference database is estimated using the log score value (≥1.8 is considered a reliable identification). Once libraries are created these can be continuously improved by inclusion of additional reference spectra capturing potential variation between strains due to geographical difference brought about by genetic variations limited arising from limited genetic exchange/gene flow.

Here we share the methodological approach to creating MALDI TOF MS libraries for entomological samples using malaria vector species identification as an exemplary case study. Moreover, the libraries related to this publication have been deposited in public repository (https://doi.org/10.7910/DVN/VYQFNO37) for anyone with MALDI-TOF MS equipment to adapt.

Materials and methods

Results

Reference spectral library query

Out of all the mosquito analysed by MALDI-TOF MS, 1,712/1782 were correctly identified with an accuracy of 96.1% [5]. The database was particularly precise in distinguishing An. gambiae and An. funestus complex members which clustered into three distinct branches: i) An. vaneedeni, An. funestus and An. parensis clustered, ii) An. gambiae complex, and iii) An. rivulorum and An. leesoni (Fig 1).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. MSP dendrogram of MALDI-TOF MS spectra used for database creation for species identification and An. vaneedeni and An. parensis [5].

https://doi.org/10.1371/journal.pone.0330605.g001

However, a few samples had low LSV values, suspected to be due to either protein degradation, physiological status or residual blood meals in the head and thorax (Fig 2). The application of the weighted LSV approach on the An. gambiae s.l samples that had ambiguous identification during querying, resulted in correct species identification congruent with the molecular analysis results, further improving the accuracy of species identification to 97.5% [5].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. LSVs obtained after a database query with MS spectra belonging to members of An. gambiae and An. funestus complexes.

Horizontal dashed lines represent the cut-off point for reliable identification (LSV > 1.8). Abbreviations: A.U., arbitrary units, An., Anopheles, s.s, sensu stricto, LSV., Log Score value [5].

https://doi.org/10.1371/journal.pone.0330605.g002

Impact of preservation method on mosquito identification and LSV distribution

With a well curated, preservation methods do not significantly impact on mosquito identification and LSV distribution as we noted that there was no significant difference (Kruskal-Wallis test, p = 0.19) in the median LSV among different methods of preservation (Fig 3).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Log Score Value (LSVs) obtained after MSP reference database query with MS spectra based on method of preservation.

Horizontal dashed lines represent the cut-off point for reliable identification (LSV > 1.8) [5].

https://doi.org/10.1371/journal.pone.0330605.g003

Discussion

This article outlines a comprehensive procedure for utilisations of MALDI-TOF MS for malaria vector species identification. The accuracy in distinguishing between sibling species of the two Anopheles’ species complex (An. funestus s.l and An. gambiae s.l) demonstrates the discriminative power of MALDI-TOF MS. In the context of surveillance, the wide range of species representation should reflect as much as possible the diversity of mosquitoes in each study area. Different mosquito species play varied roles in disease transmission and have different feeding and resting proclivities. The composition of vectors paints a picture of the type of rigor required for effective entomological surveillance and the obvious benefit of using a highly sensitive tool such as MALDI-TOF MS. The spectra obtained from the head and thorax can simultaneously measure several entomological parameters if libraries are created for that purpose. In different studies, the head and thorax spectral pattern has been successfully used for species identification [5], pathogens infection status determination [21], endosymbionts [22] and age-grading [26]. More recently, a proof-of-concept study demonstrated that spectral patterns from the head and thorax were used for discriminating between resistant and susceptible mosquitoes [27]. Here, we propose using the head and thorax spectral for all future MALDI TOF MS spectra acquisition for malaria vector profiling. To make this technology accessible to all stakeholders, including national public health agencies and partners responsible for vector surveillance, it is essential to harmonize protocols and methodologies across research groups and institutions. Standardizing approaches will enable the development of a shared platform, ensuring consistency, comparability, and broader adoption of this technology in national malaria control programs.

In conclusion, the proposed work flow enables the creation of in-house MALDI-TOF MS libraries with relative ease (protocol and associated video are attached as supplementary files). MALDI TOF MS can be used to robustly identify mosquito species. The high accuracy and ability to distinguish complex species and visualise species relationships make this method valuable in entomological monitoring of primary Afrotropical malaria vectors and can be expanded to include other mosquito disease vectors.

Supporting information

Acknowledgments

We are grateful to the Scientific and technical teams at the Centre for Geographic Medicine Research Coast, Kilifi for help in design and implementation of this work. Many thanks to the technical and field staff (Festus Yaa, Gabriel Nzai, Robert Mwakesi and Martha Muturi) who devoted their time and assisted in the field collection of samples. This paper has been published with the permission of the Director of the Kenya Medical Research Institute (KEMRI).

Associated content

- Mosquito Database Creation Using MALDI TOF MS video demonstration available on YouTube (https://youtu.be/btS0hsx9yNw). The video shows the experimental steps of MALDI TOF MS database creation.