Adult An. gambiae s.l. were sampled from rural sites along the north shore of Lake Victoria, west of the city of Kisumu in western Kenya, as described in Huang et al.  and Odiere et al. . Collections included hand catches using a mouth aspirator from houses and from pits dug into the ground, Colombian curtain house-exit samples, indoor pyrethrum spray catches, and from clay pots set out-of-doors. Control DNA was obtained from whole mosquitoes of strains available in laboratory colonies and from the Malaria Research and Reference Reagent Resource Center (MR4), namely An. gambiae ZAN/U (MRA-594), a DDT resistant strain originating from Zanzibar; An. gambiae KISUMU, originating from western Kenya; An. gambiae RSP, a strain with reduced sensitivity to permethrin originating from western Kenya; and An. arabiensis KGB (MRA-339). Heterologous DNA was from Ochlerotatus triseriatus TOUMEY WOODS strain, originating from Michigan State University; and Aedes albopictus courtesy of Dr. Steven Juliano, originating from Florida, USA. Colony conditions were as in Benedict  and Huang et al. . Mosquito samples were stored individually in 1.5 ml snap-top Eppendorf tubes at -20°C, and mosquito DNA was prepared for PCR as in Rafferty et al , with exceptions as outlined below.
Plasmid DNA controls
Plasmid control papers were prepared for the identification of An. gambiae and An. arabiensis when using the protocol of Scott et al. , and were used in TaqMan genotyping here as well. The ZAN/U (MRA-594) strain of An. gambiae and the KGB (MRA-339) of An. arabiensis were used. DNA from individual mosquitoes was prepared for PCR by the method of Rafferty et al. . A portion of the IGS including bases 177–981 (Genbank U10135) was amplified from both species using primers 5'-CCTAACAACCCTCTGAGATCC-3' and 5'-CATGCACAAGACATCCTACTACC-3'. PCR reactions consisted of 1 U of Taq DNA polymerase (Promega), 0.3 mM MgCl2, primers at 1 μM each, 0.08 mM dNTPs, and the polymerase manufacturer's suggested buffer in 25 μl total volume. PCR was performed using a Bio-Rad iCycler (Bio-Rad Life Sciences Research, Hercules, California, USA) using the following conditions: 95°C 5 m, 30 cycles (95°C 30 s, 50°C 30 s, 72°C 30 s), 72°C 5 m. The 805 bp fragment was cloned using the pGEM T-Easy Vector kit (Promega, Madison, Wisconsin, USA) according to the manufacturer's recommendations and transformed into JM109 cells. Clones were sequenced and matched those of Genbank U10135 (An. gambiae) and AF470100 (An. arabiensis). Plasmids were purified using the Qiagen-tip 500 according to the manufacturer's instructions. Strips of Schliecher & Schuell 903 cards (Whatman, Florham Park, New Jersey, USA) were dipped in water containing 0.5 ng/μl each of both plasmids, allowed to dry overnight, and stored over silica gel at room temperature. Paper spots were punched using a Harris Micro Punch 2.0 mm (Whatman) and placed directly in the PCR reaction mix as needed.
Conditions for standard and quantitative TaqMan PCR
Individual mosquito specimens from field collections or laboratory strains were prepared for identification by removing a leg with sterile forceps and placing it into one well of a 96-well PCR tray (P/N 951020389 Brinkmann Instruments, Inc., Westbury, New York, USA). Each well contained 40 μl of TE buffer (10 mM Tris-HCl/1 mM EDTA pH 8.0). Trays were covered securely with sterile adhesive foil and placed on water in a sonicator bath (Bransonic ultrasonic cleaner, Shelton, Connecticut, USA) for 5 min. For some samples, 1 μl was taken for conventional PCR using the method discussed in Scott et. al. . All conventional PCR reactions were performed using the Epicentre FailSafe PCR System (Epicentre Biotechnologies, Madison, Wisconsin, USA). The conditions for conventional PCR consisted of 25 μl of 2X Premix E, 1 μl (20 ng) of universal forward primer (5' GTGTGCCCCTTCCTCGATGT), 1 μl (12 ng) of An. gambiae s.s. specific reverse primer (5' CTGGTTTGGTCGGCACGTTT), 1 μl (20 ng) of An. arabiensis reverse primer (5' AAGTGTCCTTCTCCATCCTA), 1 μl of Epicentre PCR enzyme mix, and sufficient DNA grade water to raise the reaction volume to 50 μl. The reaction programme had an initial step of 80°C for 1 min, followed by 30 cycles of denaturation at 94°C for 30 sec, annealing at 50°C for 30 sec, and extension at 72°C for 30 sec, with a final extension at 72°C for 4 min. The PCR products were separated by electrophoresis on 2% agarose TBE gels, and stained with ethidium bromide. The amplicons were visualized with an ultraviolet transillumination gel documentation system (AlphaImager 2200, San Leandro, California, USA). The predicted DNA bands on the gel (390 bp for An. gambiae, 315 bp for An. arabiensis) were compared to a 1 Kb reference ladder.
The original tray containing the remaining mosquito sonicates was processed for TaqMan Genotyping Analysis on an ABI Model 7900 HT workstation (Applied Biosystems, Foster City, California, USA). The reaction used TaqMan mastermix (Applied Biosystems P/N 4304437) and the following newly designed Anopheles gambiae s.l. universal primer sequences: (Forward) 5'-GTGAAGCTTGGTGCGTGCT-3' and (Reverse) 5'-GCACGCCGACAAGCTCA-3'. These primers correspond to the 623–641 and 772–788 positions of the 5' end of the intergenic spacer region, respectively. A set of two species-specific TaqMan probes, conjugated to minimum binding groove ligands and a quencher at the 3' end, were synthesized to detect "allele X" (An. gambiae): 5'VIC-CGGTATGGAGCGGGACACGTA-3' and "allele Y" (An. arabiensis): 5' 6FAM-TAGGATGGAGAAGGACACTTA-3'. These probes correspond to positions 744–764 of the 5' intergenic spacer region, where the consensus sequence is TGGTATGGAGCGGGACACGTA . In this region, An. gambiae s.s. has two nucleotide substitutions, at positions 744 and 757, whilst An. arabiensis has five substitutions, at positions 745, 747, 754, 755, and 762. There are no insertions or deletions for either species in the 744–764 part of this sequence. Results were expressed as fluorescence intensity, and were displayed on an X-Y bivariate plot. The probability that the probe bound to complementary DNA of the homologous, species-specific amplicon is established with a proprietary algorithm, incorporated into the Applied Biosystems Sequence Detector v1.7 software used here, that invokes a maximum likelihood estimator and cluster analysis  to call the alleles. Occasionally, sequences were called manually by visual inspection of the fluorescence values, when automatic detection by system software failed to assign a sequence and an undetermined result was returned.
In the first experiment, a set of known sources of mosquito DNA from laboratory strains as described above, or from field specimens reliably determined previously to species by standard PCR , were used along with the plasmid controls discussed above, as well as blank well and internal negative controls. Standard PCR was repeated alongside the TaqMan genotyping method. In the second experiment, a random sample An. gambiae s.l. collected from houses in a rural area west of Kisumu, in western Kenya  was analysed by TaqMan genotyping and standard PCR. For experiments 1 and 2, the TaqMan genotyping method was analysed for sensitivity using an on-line clinical calculator  with the following equation: sensitivity = (number of true positives)/(number of true positives + number of false negatives). The reaction success rate, and comparative sensitivity of TaqMan genotyping to detect a true An. gambiae or true An. arabiensis sample, was compared by the test of the equality of two percentages . In the third experiment, a sample of 3,041 adult An. gambiae s.l. from the Bondo and Kombewa Districts of western Kenya  was analysed by TaqMan genotyping to assess the success rate of the reaction, and the proportion of the two species in the study population.