Phenotypic characteristics are most commonly employed for strain discrimination due to ease and cost efficiency. Knowledge of how strains normally behave (larval resting postures, growth rate, size, feeding preferences) serves as an ongoing incomplete authentication process. Observation is also an important part of colony health maintenance as observable factors such as slow growth are a quick indicator of unhealthy conditions or potential contamination. For both reasons, regular monitoring of all strains being housed should be performed by trained personnel, and any unexpected behavioral observations should be evaluated with authentication testing. Nitzki et al. [5
], for example, showed practical application of phenotypic markers alerting of genotypic introgression between laboratory populations.
Fixed phenotypic characteristics such as insecticide resistance or visible traits can be used to easily, inexpensively, and routinely scan for contamination events if available in the strains. For example, larvae of An. gambiae
often express visible characteristics such as collarless
, and black diamond
] (Figure ) which may be unique to certain strains. These same characteristics can be selected out of a polymorphic group as a strain-specific marker [19
]. It is important to note that deviations in phenotypic expression are not always a result of inter-strain contamination but can be caused by changes in the rearing environment, for example. Also, the expression of some traits can be variable making it difficult to use for routine authentication. For example, the collarless
trait can be weakly expressed, making it more difficult to confirm. For these reasons, using genetic markers in concert with phenotypic markers is advisable where possible. Detailed information about the phenotypic characteristics of An. gambiae
at the MR4 is not presented here but was considered in the development of these methods.
Routine challenge of an insecticide resistant strain removes any potential susceptible contaminants and maintains the strain's insecticide resistance levels which is of particular interest to our laboratory's research apart from authentication methods making it a very useful tool since we would be conducting this selection process anyway. Exposing larvae to an insecticide is an advantageous method of treatment as the concentration is more uniform in solution and all individuals remain in constant contact with the insecticide eliminating the threat of individual underexposure. Resistant populations should always be tested concomitantly with susceptible populations as a control, and if susceptible strains are reared alongside resistant populations, their susceptible nature should be monitored as well.
PCR was our chosen method for genetic discrimination because it can be standardized and reproduced in most laboratories, and it was, therefore, favored over more complicated (Luminex) or technical (karyotyping/allozyme) methods evaluated.
In many animal groups, the use of isogenic lines, or clones, are employed to overcome inter-laboratory variation since individuals tested are genetically homogenous. However the development of these lines involves sibling inter-mating resulting in inbreeding depression [2
]. The deleterious effects on longevity and fecundity are undesirable in routine colony maintenance [20
]. Standard lines, including phenotypically unique strains as well as "wild type" strains, are more often chosen due to their ease in maintenance and higher fitness levels. In higher orders, there is some debate of whether research employing standard lines is valid due to variable genotypes [3
]. This is also of concern for culicids [21
]. However, with alternate technologies such as microarray expression analyses, genetic function can be directly linked regardless of knowledge of the individual's genotype [23
Hii et al. [6
] and Dejong et al. [24
] reported inter-laboratory genetic variability in sub-populations of anopheline strains. Although these differences may be due to genetic drift and selection, contamination cannot be initially excluded. Since genetic contamination can occur without complete mixing of the chromosomes, an assay only detecting one marker on a single chromosome could miss this type of contamination. Therefore, assays should employ multiple markers on multiple chromosomes where possible for a higher level of confidence in results. In the example presented in this paper, we chose molecular markers located on the X, 2L, and 3L chromosomes.
We analyzed Luminex technology as an alternative method for authentication. With this technology, large numbers of SNP sites can be simultaneously detected in a gene(s) [25
]. By multiplexing a PCR with several genes of interest, we were able to create a successful Luminex procedure for screening samples for authentication (data not presented). However, the PCR plus Luminex procedure was more time consuming and expensive than the 1 or 2 PCR methods required as presented here. If more cost- or time-efficient, Luminex technology could be a valuable authentication tool, assuming adequate SNP sites were available for discrimination. The different genes of interest would need to be limited as well to avoid many preliminary PCR steps.
Allozyme analyses and karyotyping are successfully used to discriminate between conspecific laboratory strains [6
]. These methods, however, require extensive training in sample preparation and analyses making routine usage difficult, especially if simpler, less expensive methods are available.
We also considered insertion patterns of transposable elements (TEs) (i.e. mobile genetic units capable of replicating and spreading in the host genome) which have recently been applied to study genetic differentiation between An. gambiae
molecular forms [28
]. Among TEs, Short INterspersed Elements (SINEs) have been extensively used as phylogenetic and population genetic markers in primate taxa and An. gambiae
]. Despite the discriminative value for our strains detected in a few loci during preliminary research (data not presented), we did not pursue this approach for a number or reasons. Most loci tested were polymorphic in most strains and therefore only a few presented fixed variants with discriminative value. Moreover, most loci presented a high percentage of PCR failure (6-49%). For routine screenings, this was especially important as the absence of any amplification product for a particular sample could indicate the presence of a particular mutation in the primer binding sequence or simply a PCR failure. Also, using this method, common "stutter" bands and shadow bands made routine scoring quite difficult. In addition, the few strains that could be classified with this approach could also be authenticated using other techniques with greater reliability.