Non-random transmission of parental alleles into crop-wild and crop-weed hybrid lineages separated by a transgene and neutral identifiers in rice

It is essential to assess environmental impact of transgene flow from genetically engineered crops to their wild or weedy relatives before commercialization. Measuring comparative trials of fitness in the transgene-flow-resulted hybrids plays the key role in the assessment, where the segregated isogenic hybrid lineages/subpopulations with or without a transgene of the same genomic background are involved. Here, we report substantial genomic differentiation between transgene-present and -absent lineages (F(2)-F(3)) divided by a glyphosate-resistance transgene from a crop-wild/weed hybrid population in rice. We further confirmed that such differentiation is attributed to increased frequencies of crop-parent alleles in transgenic hybrid lineages at multiple loci across the genome, as estimated by SSR (simple sequence repeat) markers. Such preferential transmission of parental alleles was also found in equally divided crop-wild/weed hybrid lineages with or without a particular neutral SSR identifier. We conclude that selecting either a transgene or neutral marker as an identifier to create hybrid lineages will result in different genomic background of the lineages due to non-random transmission of parental alleles. Non-random allele transmission may misrepresent the outcomes of fitness effects. We therefore propose seeking other means to evaluate fitness effects of transgenes for assessing environmental impact caused by crop-to-wild/weed gene flow.