Introduction
Several species of mosquitoes are vectors of viruses or parasites causing serious diseases in humans. In particular, some species ofAnopheles transmit the Plasmodium parasite, which causes malaria in humans [1] , and thus strongly impacts human health with more than 247 million cases over 84 countries in 2021[2] . Despite global effort to reduce the burden of vector-borne diseases, they remain a sanitary and economic threat in the intertropical area and beyond [3] . However, among the 3500 species of mosquitoes, only a minority are vectors of pathogens, therefore the identification of species is an important point in surveillance and studies.
In French Guiana, 245 mosquito species are known [4,5]including 22 species of Anopheles [6] , nine of which belong to the Nyssorhynchus subgenus whose elevation to genus level is under consideration [7–9] . Among them,Anopheles darlingi is known to be the main malaria vector in South America [10,11] and Anopheles aquasalis , although not being incriminated as a vector in French Guiana, is also considered as one of the principal vectors in other neighboring countries on the continent [10,12] . Moreover,Anopheles medialis , Anopheles nuneztovari andAnopheles oswaldoi have been found naturally infected byPlasmodium in French Guiana [13] , as well asAnopheles braziliensis and Anopheles triannulatus in Brazil [14,15] , but the real extent of their involvement in parasite transmission between humans is currently unknown. Given these differences in their ability to transmit malaria, it is thus important to identify the exact species of Anopheles mosquitoes when studying and monitoring field population distributions and dynamics.
Originally, mosquito species identification was only based on visual observation of morphological characteristics with the help of dichotomous taxonomic keys [16–19] . This method is effective and relatively accessible when morphological differences are substantial, yet it is arduous and requires advanced skills when variations are subtle. The use of additional equipment and the dissection of internal structures are also sometimes mandatory to distinguish between morphologically close species, as for Culexmosquitoes, in which meticulous dissection of male genitalia is required[20,21] . Moreover, even with properly trained and experienced professionals the task frequently remains a challenge, misidentifications are common and results for later processed samples cannot be checked afterward in case of doubt [22] .
Alternative methods have been developed based on molecular biology techniques. CO1 (Cytochrome C Oxidase Subunit 1), a mitochondrial gene, and ITS2 (Internal Transcribed Spacer 2), a non-coding nuclear sequence located between 5.8S and 28S ribosomal RNA genes, are often used for molecular identification of animals. CO1 or ITS2 DNA barcoding are two of the most prevalent techniques for species identification today, with databases usually already available. Amplification of DNA regions with species-specific sizes via multiplex PCR (Polymerase Chain Reaction) is often used as a substitute to sequencing to discriminate a limited number of species found locally[23–26] , notably Anopheles coluzzii andAnopheles gambiae found in Africa [27,28] and other species from the Gambiae Complex [29,30] . When a higher number of species are involved, PCR may be combined with a treatment with restriction enzymes in the RFLP (Restriction Fragment Length Polymorphism) technique [31,32] .
Other approaches for mosquito species identification are currently being considered. Mosquito protein profiling using MALDI-TOF technology is promising [33] , even though it requires a substantial investment in terms of equipment and the databases are still under development [34] . Nowadays, increasingly powerful software solutions make it possible to perform morphometry of mosquito wings[35] and the use of artificial intelligence may allow for automatic visual identification of mosquitoes [36] . The sound and frequency of wing beats is also used to develop identification tools that are meant to be accessible on a simple smartphone[37] . Yet the implementation of these tools needs more development and advanced computational knowledge.
In this study, we have developed a new method to distinguish between nine Anopheles species from French Guiana based on precise discrimination of their natural ITS2 sequence size polymorphism using capillary electrophoresis, combined with a simple morphological observation, the color of the hindleg tip (fifth hind tarsus, Ta-III5).