Contemporary gene flow between wild An. gambiae s.s. and An. arabiensis

Background: in areas where the morphologically indistinguishable malaria mosquitoes anopheles gambiae giles and an. arabiensis patton are sympatric,hybrids are detected occasionally via species-diagnostic molecular assays. an. gambiae and an. arabiensis exhibit both pre- and post-reproductive mating barriers,with swarms largely species-specific and male f1 (first-generation) hybrids sterile. consequently advanced-stage hybrids (back-crosses to parental species),which would represent a route for potentially-adaptive introgression,are expected to be very rare in natural populations. yet the use of one or two physically linked single-locus diagnostic assays renders them indistinguishable from f1 hybrids and levels of interspecific gene flow are unknown. methods. we used data from over 350 polymorphic autosomal snps to investigate post f1 gene flow via patterns of genomic admixture between an. gambiae and an. arabiensis from eastern uganda. simulations were used to investigate the statistical power to detect hybrids with different levels of crossing and to identify the hybrid category significantly admixed genotypes could represent. results: a range of admixture proportions were detected for 11 field-collected hybrids identified via single-locus species-diagnostic pcrs. comparison of admixture data with simulations indicated that at least seven of these hybrids were advanced generation crosses,with backcrosses to each species identified. in addition,of 36 individuals typing as an. gambiae or an. arabiensis that exhibited outlying admixture proportions,ten were identified as significantly mixed backcrosses,and at least four of these were second or third generation crosses. conclusions: our results show that hybrids detected using standard diagnostics will often be hybrid generations beyond f1,and that in our study area around 5% (95% confidence intervals 3%-9%) of apparently 'pure' species samples may also be backcrosses. this is likely an underestimate because of rapidly-declining detection power beyond the first two backcross generations. post-f1 gene flow occurs at a far from inconsequential rate between an. gambiae and an. arabiensis,and,especially for traits under strong selection,could readily lead to adaptive introgression of genetic variants relevant for vector control. © 2014 weetman et al. licensee biomed central ltd.