The scope of Rp EPIC markers in population genetic studies: a preliminary study with dengue vectors

This article is published under the Creative Commons CC-BY-ND License (http://creativecommons.org/licenses/by-nd/4.0/). This license permits use, distribution and reproduction, commercial and non-commercial, provided that the original work is properly cited and is not changed in anyway. Abstract: Most genetic markers carry numerous limitations which question their credibility. Ribosome protein Exon Primed Intron Crossing (Rp EPIC) markers claimed to eff ectively overcome some of these limitations. Hence, the present study was aimed at demonstrating the scope of Rp EPIC markers in population genetics in comparison to conventional microsatellites, using the dengue vectors; Aedes aegypti and Aedes albopictus. A total of 62 Ae. aegypti and 51 Ae. albopictus samples were genotyped for two Rp EPIC markers; RpS20b and RpL30a and two microsatellite markers; AC7 and BbH08 for Ae. aegypti and Alb-tri 03 & Alb-tri 25 for Ae. albopictus. The Rp EPIC markers were successfully transferred among the two species, while diff erentiating them based on F statistics. Total size variants (electromorphs) observed for the two markers, RpS20b and RpL30a were 14 (3 for Ae. aegypti; 13 for Ae. albopictus; shared: 2) and 7 (3 for Ae. aegypti; 4 for Ae. albopictus; shared: 0) with many more alleles uncovered through sequencing (RpS20b:74 and RpL30a:54). In contrast, microsatellites produced 6 (AC7) and 3 (BbH08) size variants for Ae. aegypti and 8 (Alb-tri 03) and 7 (Alb-tri 25) for Ae. albopictus. Both marker types detected genetic structure among Ae. albopictus populations while a genetic structure among Ae. aegypti was detected only with microsatellites and not with Rp EPIC size variants. The combined data of microsatellites and Rp EPIC size variants as well as sequence analysis did not support the presence of a genetic structure among the studied Ae. aegypti populations. It is possible that microsatellites tend to infl ate genetic diff erentiation among the studied Ae. aegypti populations, which has been counteracted by the Rp EPIC markers.


INTRODUCTION
During the last fi fteen years, use of molecular markers played a major role in understanding the genetic diversity and evolution within living organisms revealing polymorphisms at the DNA level. Selecting an appropriate set of genetic markers, which best answer the target questions would be the fi rst and most critical step in any such study. There are several types of genetic markers, which have been used in population genetic studies of insect vectors and other organisms by numerous scientists with both pros and cons in terms of cost, speed, amount of DNA needed, technical labour, degrees of polymorphism, precision of genetic distance estimates, and statistical power of tests. For example, Restriction Fragment Length Polymorphism (RFLP) markers have been used in analysing population genetic aspects of insect vectors such as Ae. aegypti (Yan et al., 1998) and Anopheles gambiae (Romans et al., 1991) which are expensive, time consuming, less polymorphic and require high quality DNA (Francis et al., 2017;Singh et al., 2017). The Amplifi ed Fragment Length Polymorphism (AFLP) technique was found to be more advantageous than RFLP with improved specifi city, reproducibility, abundance on the genome, ability to develop without prior knowledge of the sequence and fl exibility to use on both dead and live specimens (Gupta & Preet, 2014;Singh et al., 2017). However, they are expensive, more laborious, time consuming and are of dominant nature ( Yan et al., 1999;Francis et al., 2017).

September 2020
Journal of the National Science Foundation of Sri Lanka 48(3) In late 1900's, the development of Polymerase Chain Reaction (PCR) technique led to amazing advances, mainly in research of genetic diversity. Randomly Amplifi ed Polymorphic DNA (RAPD) was the fi rst PCR based technique which was simpler to use, easy to develop and capable of capturing high resolution of diversity across large regions of the genome at one time, but the lack of reproducibility between diff erent runs due to short primer length and dominant nature aff ected its credibility (Apostol et al., 1996;Francis et al., 2017).
With the advent of sequencing techniques, mitochondrial DNA (mtDNA) became the marker of choice for population genetic studies of various organisms including insect vectors such as dengue and malaria species (Yang et al., 2011;Twerdochlib et al., 2012;Tchouassi et al., 2014;Weeraratne et al., 2018;Motoki et al., 2019). Use of this marker in population genetics seemed to be advantageous as they were easy to amplify, sequence and analyse, maternally inherited, and high in evolutionary rate with low eff ective population size although it carried some limitations such as integration with the nuclear genome, and non-neutral evolution etc. (Hurst & Jiggins, 2005;Francis et al., 2017). Even the most widely used genetic marker type in the last decade, simple-sequence repeats (SSR, also known as microsatellites) have shown to carry several limitations when used alone; viz. low genomic frequency, null alleles, band stuttering and low polymorphisms (Fagerberg et al., 2001;Janderson et al., 2002;Goubert et al., 2016;Wei et al., 2019). Besides these pros and cons, these markers have been used in population genetic studies covering a large spectrum of fauna and fl ora in Sri Lanka (Gunawickrama, 2007;Dammannagoda et al., 2008;De Croos & Pálsson, 2010;Rajkumar et al., 2011;Abeysinghe et al., 2014;Dangalle et al., 2015;Sandamal et al., 2018;Weeraratne et al., 2018). The Ribosomal protein Exon Primed Intron Crossing (Rp EPIC) markers, which have been used to infer genetic variability since 1990's, fl ank the exon regions and detect the polymorphisms across the conserved intron region (Lessa, 1992;Slade et al., 1993). Although they were primarily used to analyse the natural history of species, more recent studies have proved their usefulness in detecting population genetic structures. Most importantly, they have shown to be transferable among species with the use of same PCR primers and even with the same PCR profi les (White et al., 2015). They were used in population genetic studies of many organisms across various phyla. For example, Touriya et al. (2003) successfully used Rp EPIC markers in population genetic studies in marine and freshwater fi sh species and were able to assess the universality of primers by testing them on both reptiles (Moorish gecko) and mammals (human and camel). Eff ectiveness of Rp EPIC markers with invertebrates was demonstrated in a population genetic and phylogeographical study of six ant species in Brazil (Ströher et al., 2013). Rp EPIC markers have also been successfully used for pests, i.e. to detect the population substructure at temporal and spatial scales in cotton bollworm Helicoverpa armigera in India, Australia and China (Tay et al., 2008) and for disease vectors, i.e. to diff erentiate dengue vector Ae. aegypti from Australia, parts of Thailand and Vietnam (Endersby et al., 2009;Endersby et al., 2011). It has also been shown that Rp EPIC markers are transferable among related vector species, i.e. same Rp EPIC markers can be used to detect the population structure of Ae. aegypti, Ae. albopictus and Ae. notoscriptus indicating its versatility in vector genetics (White et al., 2015).
However, there are no reports available on studies conducted using Rp EPIC markers in Sri Lankan context. Hence, the aim of the present study was to examine the scope and usefulness of Rp EPIC markers in population genetic studies using dengue vector mosquitoes in Sri Lanka. The rationale for selecting dengue vectors to study the scope of Rp EPIC markers in population genetics is explained below in the context of dengue infection in Sri Lanka.
Sri Lanka is a country which had invested immensely in its struggle to combat dengue outbreaks, since its hyper-endemic situations in 2004 with 15,463 suspected dengue cases and 88 deaths (National Plan of Action for Prevention and Control of Dengue Fever 2005-2009). In 2017, the number of dengue cases reported in the country was 186,101 while the death toll was more than 400 as per the records from the National Dengue Control Unit (NDCU), which is the highest seen in recent years. It is nearly a 12-fold increase in dengue cases compared to the fi rst dengue epidemic in the country in 2004. Hence, dengue infection has become a major contributing factor that declines the health index in Sri Lanka, which necessitates immediate attention by all the respective parties in order to achieve an eff ective disease control method. Although the World Health Organization (WHO) has endorsed the new dengue vaccine, Dengvaxia, which is anticipated to be 70 % eff ective, to be used in the countries that are battling with dengue, the clinical trials are still continuing including in Sri Lanka. However, the feasibility of using the vaccine in controlling the dengue transmission is still unclear. Hence, vector elimination still remains as the major control and preventive method of dengue infection in countries batting with dengue epidemics. Although the massive vector control campaigns conducted by various governmental authorities soon after epidemic situations had been successful in reducing the number of dengue cases time-to-time for short periods, the eff orts towards a  White et al. (2015). Further, both markers produced PCR amplicons within the size range that can be separated on polyacrylamide gels which was an added advantage.
The results were compared with the data generated with two comparable microsatellite markers for each species and also with those data generated by combining both the Rp EPIC and microsatellite markers to understand the relative role of Rp EPIC markers in the given set-up. The four microsatellite markers used in the study were AC7 (Slotman et al., 2007) and BbH08 (Chambers et al., 2007) for Ae. aegypti and Alb-tri 03 and Alb-tri 25 (Beebe et al., 2013) for Ae. albopictus. The two microsatellite markers, AC7 and BbH08 have been utilised for many population genetic studies of Ae. aegypti in the world. For example, AC7 has been used in investigating the population structure of Ae. aegypti in Thailand and Kenya (Slotman et al., 2007) and mainland Southeast Asian countries (Hlaing et al., 2010) whereas BbH08 has been used in Trinidad (Chambers et al., 2007) and Thailand (Olanratmanee et al., 2013). In general, these markers have provided additional tools to understand the population genetic structure and gene fl ow within Ae. aegypti populations facilitating the identifi cation of the patterns of disease transmission (Slotman et al., 2007;Hlaing et al., 2010;Olanratmanee et al., 2013). The two microsatellites that amplifi ed Ae. albopictus in this study, Alb-tri 03 and Alb-tri 25 have been fi rst used in a population genetic study conducted in the Australasian region and were found to carry suffi cient diversity to detect population diff erentiation among Ae. albopictus populations (Beebe et al., 2013). Irrespective of the inherent drawbacks of microsatellites, these two markers have proved to be eff ective in population genetics in many instances (Beebe et al., 2013;Minard et al., 2018).

METHODOLOGY
Ethical clearance for this study, including the sampling of mosquitoes, was granted by Ethical Review Committee of Institute of Biology, Sri Lanka (ERC IOBSL 122 04 15).

Sample collection
Mosquito samples were collected from areas where each species showed a high prevalence at the time of collection (May -October 2015) based on the information gathered from the Dengue Epidemiology Unit, Ministry of Health. In each study area, an average of 100 ovitraps (ranging from 90 to 110) was placed covering an area of 500 m in diameter approximately. Although the initial aim was to collect both species from the same sampling areas, this was not possible as adequate numbers of samples of both species could not be collected at the same site. For example, Ae. aegypti was sampled from two urban areas (Colombo city area; number of samples, N = 32) and Galle Fort; N = 30). However, the number of Ae. albopictus larvae collected from Colombo and Galle Fort was not suffi cient (14 and 11, respectively) for a population genetic study. Accordingly, Ae. albopictus was sampled from two suburban areas due to its reputation to prevail more in suburban and rural areas: Kalamulla (N = 32) in Kalutara District and Imbulgoda (N = 19) in Gampaha District. All four study areas suff er frequent dengue outbreaks due to many reasons. Colombo being the commercial capital of Sri Lanka, encompasses the central hubs of public transportation and all sorts of working places, factories, schools, and shopping malls within the city limits, which cause thousands of private vehicles to enter the city every day from all parts of the country. This endless public and private transportation mainly acts as a good source of vector transfer in and out of the city. Moreover, rapid urbanisation along with poor city sanitation, inappropriate waste disposal and failure to conduct continuous vector removal make home for mosquito breeding. Galle is another urbanised busy city where tourism and fi shing industry prevail. Due to the fl ow of visitors on regular basis coupled with a poor garbage removal system in place, the area provides opportunity to create plenty of mosquito breeding sites and transfer of mosquitoes. On the other hand, Kalamulla and Imbulgoda are residential areas with ample vegetation including home gardens, paddy fi elds etc. and are located close to towns and commercial centres.

September 2020
Journal of the National Science Foundation of Sri Lanka 48 (3) All the ovitraps used for mosquito collection were prepared according to WHO guidelines (WHO, 2009). Each ovitrap consisted of a 250 mL black painted plastic container of 6.5 cm in diameter and 12.0 cm in height. A stripe of fi lter paper (29.7 cm × 7 cm) was laid around the inner surface of each container and it was fi lled with hay infusion up to about 6.0 cm level. Hay infusion was prepared previously by steeping 125 g of dry rice hay in 120 L of dechlorinated water for 7 d in a tightly closed plastic bucket. Ovitraps were placed at indoor and outdoor sites in houses in the selected areas choosing spots protected from rain, direct sunlight and beyond reach of children and pets. The ovitraps were collected 5 d after placement and both eggs and larvae were transferred to the laboratory in glass bottles containing the same hay infusion. Ovitrap placements and collections were done after obtaining the informed consent from the occupants.
Once transferred to the laboratory, eggs were kept for hatching and the larvae were reared up to fourth instar level. At the fourth stage, larvae were killed by direct submergence in absolute ethanol and identifi ed to species level using standard taxonomic keys (Amerasinghe, 1995). Identifi ed Ae. aegypti and Ae. albopictus larvae were preserved in absolute ethanol and stored at -20 0 C prior to DNA extraction and genetic analysis. In order to avoid incidental sampling of closely related individual mosquitoes, one mosquito larva was randomly selected from each positive ovitrap for the DNA extraction.

Extraction of DNA
Mosquito larvae preserved in absolute alcohol were air dried and transferred to separate 1.5 mL microcentrifuge tubes. To ensure complete evaporation of ethanol, larvae were again dried in a 55 0 C dry bath for 20 min and each larva was crushed using a micropestle to make a fi ne powder. DNA was extracted with Promega Wizard genomic DNA purifi cation kit, following manufacturer's instructions. The resulting DNA pellet was air dried for approximately 30 min and resuspended in 100 µL of 10 mMTris-1 mM EDTA (pH 8.0). All DNA samples were stored at 4 0 C for short term and at -20 0 C for the long term.

Generation of Rp EPIC markers
Since this is only a pilot study, which investigates the feasibility of using Rp EPIC markers in the analysis of population genetics in Ae. aegypti and Ae. albopictus mosquito populations in Sri Lanka, only two markers: RpL30a and RpS20b (Endersby et al., 2009) were used.
Polymerase chain reaction (PCR) amplifi cation of the markers were carried out according to Endersby et al. (2009) with some modifi cations. Primer sequences for the two markers were RpL30a -Forward primer: ATGGTTACCGCCAAGAAACA, Reverse primer: CGGAGAGTCTTCAGGGTCTG; RpS20b -Forward primer: GCGTATYACCACCCGTAAGA, Reverse primer: GCGAGTGCARATCGATGATA. Each PCR reaction was performed in a 25 µL volume comprising 25 ng of extracted DNA, 1X PCR amplifi cation buff er, 1.5 mmol/L MgCl 2 (Promega, USA), , 0.2 mmol/L dNTPs (Promega, USA), 0.1 µmol/L forward primer (IDT, USA) and 0.1 µmol/L reverse primer (IDT, USA) for RpS20b, 0.15 µmol/L forward primer (IDT, USA) and 0.15 µmol/L reverse primer (IDT, USA) for RpL30a and 1.75 U of Go Taq DNA Polymerase (Promega, USA). The same PCR cycling conditions were used for both Rp EPIC markers for both species: initial denaturation (5 min, 94 0 C), 35 cycles of denaturation (30 s) at 94 0 C, annealing (30 s) at 58 0 C and extension (30 s) at 72 °C, fi nal extension at 72 °C (10 min). Fragments derived from PCR were fi rst run on 2 % agarose gels for qualitative analysis and then separated on 6 % denaturing polyacrylamide gels (PAGE) followed by silver staining to determine the zygosity. The alleles were scored manually based on their sizes using known samples, which were confi rmed previously through automated Sanger sequencing.

Processing of heterozygous samples
All heterozygous samples detected from PAGE were purifi ed to obtain the individual alleles by incising the gel before sending them for sequencing. Initially, the two alleles resulting from any given sample were cut apart from the polyacrylamide gel using a sterile surgical blade and put into separate 1.5 µL microcentrifuge tubes. A volume of 50 µL of 10 mMTris-1 mM EDTA (pH 8.0) was added to each tube and they were incubated at 37 0 C overnight to elute DNA from the gel. A second PCR amplifi cation was performed using a volume of 3 µL from each eluted DNA sample with the same primers and using the same PCR conditions. PCR amplicons were again run on 2 % agarose and 6 % denaturing polyacrylamide gels to confi rm the accuracy of allele purifi cation.

Sanger sequencing
PCR amplicons of all homozygous samples and purifi ed Rp EPIC alleles of the heterozygous samples were sent to Macrogen, South Korea for automated Sanger sequencing.

Journal of the National Science Foundation of Sri Lanka 48(3)
September 2020

Generation of microsatellite markers
In order to compare the usefulness of Rp EPIC markers with microsatellite markers in determining population structures of Ae. aegypti and Ae. albopictus, two established microsatellite markers for each species were generated; AC7 and BbH08 for Ae. aegypti and Alb-tri 03 and Alb-tri 25 for Ae. albopictus. For each marker, the 25 µL PCR reaction mix comprised 25 ng template DNA, 0.2mmol/L dNTPs (Promega, USA), 2.5 µL of 10X Dream Taq

Data analysis
For both Rp EPIC markers, alleles generated using each method, i.e. with polyacrylamide gel electrophoresis (PAGE) and sequencing were analysed separately. For each marker detected in PAGE (both Rp EPIC and microsatellites), allele numbers, allelic richness, Weir and Cockerham's measure of F IS and overall F ST estimates were calculated using FSTAT version 2.9.3 (Goudet, 1995). Estimates of observed (H O ) and expected (H E ) heterozygosities, conformity to Hardy-Weinberg equilibrium (HWE), the linkage disequilibrium (LD) and population pairwise genetic diff erentiation (F ST ) was estimated in Arlequin version 3.01 (Excoffi er & Liscer, 2006). The level of gene fl ow between the populations within the species was calculated following the formula Nm = 1/4 (1/ F ST -1) (Slatkin & Barton, 1989), where N is the eff ective population size and m is the migration rate. The null alleles and allele dropouts were determined using MICROCHECKER version 2.2.3 (Van Oosterhout et al., 2004).
Following analyses were conducted using the allele sequences generated from Rp EPIC markers. All the sequences generated for each marker and species were assembled together and aligned using BioEdit 7.2.5 (Hall, 1999), following the ClustalW multiple alignment algorithm with default parameters tovisualise single nucleotide polymorphisms (SNPs) and InDels. BLASTN search tool available at https://blast.ncbi.nlm.nih.gov/ Blast.cgi was used to compare the similarity of the generated sequences to reference genomes of Ae. aegypti and Ae. albopictus.
The Rp EPIC marker DNA sequence variability was analysed in terms of the following parameters using DnaSP version 5.10.01 (Librado & Rozas, 2009): nucleotide diversity (Pi and Watterson's θ w ), number of haplotypes, haplotype diversity and the average number of nucleotide diff erences. In addition, InDel polymorphism of each population of Ae. aegypti and Ae. albopictus was calculated in terms of the total number of InDel sites and events analysed, average InDel length, the number of InDel haplotypes, InDel haplotype diversity, InDel diversity and InDel diversity per site under the multiallelic model which considers all InDel events.

Rp EPIC marker amplifi cation and polymorphism
The two Rp EPIC markers, RpS20b and RpL30a were amplifi ed successfully in the samples of both Aedes species under the same PCR conditions. This diff ers to what had been reported by White et al. (2015) where diff erent annealing temperatures had to be used for the amplifi cation of markers among the two species irrespective of the transferability of the markers. However, the results of this study confi rmed that these Rp EPIC markers are not only well transferable between the two species but can be amplifi ed together with much ease, a feature that would be immensely convenient in evolutionary genetic studies.
The level of marker polymorphism (as detected by PAGE) diff ered widely among the two markers as well as among the two species. A lower allele count was found for RpL30a marker in both Ae. aegypti (Colombo: 2; Galle: 3; shared: 2) and Ae. albopictus (Kalamulla: 2; Imbulgoda: 3; shared :1). For RpS20b marker, allele counts remained low for Ae. aegypti (Colombo:3; Galle: 2; shared: 2) while for Ae. albopictus they were quite high (Kalamulla: 11; Imbulgoda: 7; shared :5) ( For both markers, allele sharing was observed between the populations of the two species (shared alleles for Ae. aegypti: 2 for each marker; Ae. albopictus: 5 and 1 for RpS20b and RpL30a, respectively). In addition, due to the transferability of the Rp EPIC markers across species, 2 of the alleles accounting for 14.3 % of the total alleles of RpS20b marker were shared by the two species as well. As such, a total of 14 variant alleles were observed for the RpS20b marker (13 alleles from Ae. albopictus, 3 alleles from Ae. aegypti and 2 alleles shared between the species). However, no shared alleles were observed for RpL30a marker (total number of variant alleles: 7). When the number of alleles observed in this study is compared with the reports from elsewhere, it showed that Sri Lankan Ae. albopictus populations are much more diverse with respect to RpS20b marker compared to those found in Spain and Indonesia (3 alleles each) (White et al., 2015). However, Ae. aegypti populations of the current study showed 3 alleles for the same marker compared to a similar study where 6 alleles had been reported for Ae. aegypti populations collected from three countries: Australia and some parts of Thailand and Vietnam (Endersby et al., 2009). Given the fact that larger countries harbour higher genetic diversity in general, this observation is suggestive of a relatively larger genetic diversity for the Sri Lankan mosquito population in relation to its small geographical size compared to much bigger countries. Similarly, there are other reports of comparable or lower allelic variation for the studied Rp EPIC markers (2 alleles each) in both fi eld caught and laboratory reared strains of Ae. aegypti (Yeap et al., 2011). Further, the observed diff erence in allelic diversity between local Ae. aegypti and Ae. albopictus populations might be stemming from the rapid growth and distribution of Ae. albopictus populations as a result of its high invasive nature and successful adaptability to the environmental changes (Bonizzoni et al., 2013;Waldock et al., 2013).
When the two markers were tested for the conformity to Hardy-Weinberg equilibrium (HWE), three of the eight tests showed deviations from the HWE: RpS20b for Ae. albopictus from Kalamulla and Imbulgoda and RpL30a for Ae. albopictus from Kalamulla (Table 1). Since analysis with MICROCHECKER indicated the presence of null alleles in the respective populations, the observed deviations from HWE may have resulted from these null allele problems. However, there were no large allele dropouts and linkage disequilibrium (p > 0.05 for both populations together and separately) with respect to the studied markers and populations.

Journal of the National Science Foundation of Sri Lanka 48(3) September 2020
Inbreeding coeffi cients detected for those populations which did not confer to HWE were relatively high and ranged from 0.350 (Ae. aegypti in Colombo for RpS20b) to 1.000 (Ae. albopictus in Kalamulla for RpL30a) ( Table 1). It is interesting to note that all Ae. albopictus individuals collected from Kalamulla were homozygous for RpL30a marker with only two alleles, resulting in the maximum possible inbreeding coeffi cient. A plausible explanation for this observed homozygosity of RpL30a marker is allele fi xation driven by the genetic drift. However, the substantial heterozygosity levels observed for the other marker suggests a population at mutation drift equilibrium rendering this assumption unlikely. In addition, RpL30a marker showed substantial heterozygosity in the other three populations tested in the present study as well as in previous research conducted elsewhere (Endersby et al., 2009(Endersby et al., , 2011Olanratmanee et al., 2013;White et al., 2015). This excludes the possibility of a function related sequence conservation or genetic hitchhiking for the RpL30a marker. Thus, it is likely that the observed homozygosity has been caused by the limited sample number in the current study, which necessitates further analysis of the population with a larger sample size.

Rp EPIC marker-based population diff erentiation
Pairwise F ST estimates obtained for the two populations within each species were substantially low and were not statistically signifi cant (p > 0.05) indicating lack of population structure among the studied areas ( Table 2, 'below diagonal without parenthesis'). Further, as expected, when populations between the two species were compared, there was signifi cant diff erentiation (p < 0.05) among all respective populations indicating the high discriminatory capacity of these markers notwithstanding the transferability of the markers across species. The overall estimates of Nm between populations within species were higher (∞) for Ae. aegypti compared to Ae. albopictus (3.08) denoting a higher level of mosquito migration within the two Ae. aegypti populations despite their distant geographical locations.

Size homoplasy in Rp EPIC markers
For both species, sequence analysis yielded a higher allele number in comparison to the alleles scored through PAGE, exhibiting an extensive amount of size homoplasy. The highest diff erence was marked for Ae. albopictus from Kalamulla population for RpL30a marker, which was a 12.5-fold increase of allelic variation in comparison to allele scoring on PAGE. The lowest diff erence (3) was shown for Ae. aegypti from Galle for the same marker (Table 1). In general, the average allele numbers generated from sequence analysis per marker and population were 9.75 and 22.25 for Ae. aegypti and Ae. albopictus, respectively which were signifi cantly higher (p < 0.05) than the allele numbers obtained via PAGE. This indicates that sequencing of Rp EPIC markers signifi cantly increased the level of marker polymorphism compared to the results obtained through electromorphs detected on PAGE. This increased amount of polymorphism could be further utilised as separate genetic markers in population and evolutionary studies, which is an added advantage of using Rp EPIC markers. However, in more conventional size variants like microsatellites, the number of sequence variants are observed to be lesser (Viard et al., 1998).

Nucleotide and haplotype diversity of Rp EPIC alleles
When nucleotide diversity (Pi) values were calculated, Ae. albopictus samples showed the largest values for both markers; 0.0654 (Imbulgoda population) for RpS20b and 0.0249 (Kalamulla) for RpL30a in comparison to Ae. aegypti. Within the populations of each species, Pi averaged over the two Rp EPIC markers did not diff er much; for Ae. aegypti Pi for Galle and Colombo Pi: nucleotide diversity (the average number of nucleotide diff erences per site between two sequences), θ: nucleotide diversity (nucleotide diversity calculated using the number of segregating sites), H: number of haplotypes based on DNA polymorphism exclusive of InDels (percent number of haplotypes to the total number of sites analyzed), Hd: DNA polymorphism haplotype diversity and k: average number of nucleotide diff erences.

Journal of the National Science Foundation of Sri Lanka 48(3) September 2020
populations were 0.01695 and 0.01595, respectively while for Ae. albopictus, Pi for Imbulgoda and Kalamulla populations were 0.0435 and 0.0304, respectively. The same pattern was observed for nucleotide diversity (θ), which estimates the population mutation rate from number of polymorphic sites. The haplotype diversity averaged over markers (Hd) were comparable for the two populations within each species (for Ae. aegypti, Hd Colombo : 0.6405; Hd Galle : 0.71 and for Ae. albopictus, Hd Kalamulla : 0.966; Hd Imbulgoda : 0.8595), although in between species, it was signifi cantly higher for Ae. albopictus (0.9133) compared to Ae. aegypti (0.6753) ( Table 3).
The Pi values obtained for Ae. aegypti populations for both markers of this study were comparable to the Pi values that have been previously reported for Ae. aegypti from Queensland and Brazil (White et al., 2015).
Ae. albopictus revealed moderately higher or similar diversity levels to Spain and Indonesian populations for RpS20b but lower diversity values with respect to RpL30a marker. As a whole, the two local Ae. albopictus populations seemed to be more diverse than the two local Ae. aegypti populations.

InDel (Insertion-Deletion) polymorphism
In general, the number of InDel sites was highest for    (Table 5), refl ecting a possible population expansion within that population. This is further evidenced by the high haplotype diversities observed for Ae. albopictus population at Imbulgoda with 16 singleton haplotypes without any common haplotypes for the same RpS20b marker suggesting that this population was maintaining most of the new mutations while expanding with respect to that marker. This observation is consistent with the fact that Ae. albopictus is the most invasive mosquito species in the world showing an immense capacity to adapt.

Comparison of allelic data between microsatellites and Rp EPIC markers
The analysis on HWE showed that only fi ve comparisons out of eight conferred to HWE (p < 0.05) for the Rp EPIC markers, whereas for the microsatellites, all eight comparisons were in HWE (p < 0.05). However, no signifi cant linkage disequilibrium was found (p > 0.05) for both the marker types. MICROCHECKER could fi nd null alleles for the two Ae. albopictus populations for both Rp EPIC and microsatellite markers but not in Ae. aegypti populations for either of the markers. Nevertheless, both species did not show large allele dropouts for either Rp EPIC or microsatellite marker types. These results refl ect the general suitability of both the marker types in evolutionary studies.
The number of size variants shown by microsatellites for the two species were 6 (AC7) and 3 (BbH08) for Ae. aegypti and 8 (Alb-tri 03) and 7 (Alb-tri 25) for Ae. albopictus. When the genetic diversity indices are compared among the two marker types, microsatellites tend to exhibit a larger diversity compared to the Rp EPIC markers, especially for Ae. aegypti (Table 6). For Ae. albopictus, this diff erence is much less prominent. This is probably due to the fact that microsatellites only capture certain intronic areas where variability is concentrated, while Rp EPIC markers screen the total Journal of the National Science Foundation of Sri Lanka 48 (3) September 2020 intron area, including those more conserved functionally signifi cant regions. Nevertheless, the eff ective number of alleles or the number of equally frequent alleles that would take to achieve a given level of gene diversity in a population, calculated per marker was comparable between the two marker types for both the species ( These results question the reliability of microsatellitebased estimates on genetic diff erentiation, which are likely to be infl ated due to its inherent larger variability. With this respect, Rp EPIC -based estimates seem to be far more reliable for two reasons. First that it captures a variability dispersed over a larger gene area which exclude the possible bias of selecting only those highly variable smaller gene segments. Secondly it allows the evolutionary question at hand to be investigated in diff erent angles to come up with more robust inferences, i.e. analysing molecular size variant data together with haplotypes and/or nucleotide diversity, etc. The same trend was observed when the global F ST values were calculated for the two marker types. For microsatellite markers, the global F ST values were higher for both Ae. aegypti (0.022) and Ae. albopictus (0.16) indicating a substantial genetic structure for the two species, as opposed to the global F ST estimates obtained based on Rp EPIC marker data generated on PAGE without sequencing (Ae. aegypti: -0.002; Ae. albopictus: 0.075), or F ST estimates yielded in combined analysis of microsatellite and Rp EPIC marker PAGE-generated data (Ae. aegypti: -0.004; Ae. albopictus: 0.089), which indicate subtle or absence of a genetic structure. The global F ST values generated based on Rp EPIC marker sequence analysis (Ae. aegypti: -0.0015; Ae. albopictus: 0.0753) was also quite similar to the results obtained through PAGE analysis of Rp EPIC markers. These observations question the reliability of marker dependent genetic inferences made on population characteristics. It may probably be more appropriate to combine several marker types with diff erent characteristics, such as the combined analysis of Rp EPIC and microsatellites as demonstrated here, to counteract the infl ation brought in by one marker, while allowing its desirable qualities to be utilised nevertheless.

CONCLUSION
Rp EPIC markers, RpL30a and RpS20b were successfully utilized in population genetic analysis of the selected Ae. aegypti and Ae. albopictus populations in Sri Lanka. These were transferable among the two-sister species and allowed for successful multiplex amplifi cation. So far, this is the fi rst study in Sri Lanka that has investigated the scope of Rp EPIC markers in molecular genetic studies. As was apparent with the sequencing of size variants, a relatively large amount of polymorphism was found with both Rp EPIC markers, although allelic diversity of RpL30a marker was relatively low compared to RpS20b among the size variants. However, both markers exhibited null alleles in some of the populations analysed. Further, as revealed by the current study, when analysed along with microsatellites, Rp EPIC markers are likely to counteract the infl ation of evolutionary genetic parameters. This might have resulted from the relative low variability of Rp EPIC markers, which analyse the entire intronic regions compared to the microsatellites that focus only on selected highly variable regions.