Abstract

The core–periphery paradigm predicts that asymmetric selection along environmental gradients drives divergent adaptation at range margins, influencing range limits and invasion dynamics. However, empirical tests spanning entire latitudinal gradients and mechanistic explanations for the genetic paradox of invasion—rapid adaptation despite limited genetic diversity—remain rare. We conducted a five-year reciprocal transplant common garden experiment on the invasive annual Amaranthus palmeri across its entire invasive range in China (21–49° N). Population fitness was quantified using asymptotic growth rates ( λ s $$ {\lambda}_{\mathrm{s}} $$ ) to test for asymmetric adaptation. We then measured key phenotypic traits and used structural equation modelling (SEM) to identify trait–fitness relationships. Finally, we integrated quantitative genetic ( Q ST $$ {Q}_{\mathrm{ST}} $$ ) and neutral genomic ( F ST $$ {F}_{\mathrm{ST}} $$ ) comparisons with gene-level analyses, including genome-wide association studies (GWAS), a genome-wide F ST $$ {F}_{\mathrm{ST}} $$ scan and expression validation, to elucidate the evolutionary mechanisms underlying adaptation. We detected pronounced latitudinal asymmetry in adaptation. Northern peripheral populations evolved local specialization through genetic shifts towards earlier flowering (2.5–18.9 days) and prolonged flowering duration (5.7–18.2 days), increasing fitness (| ß | > 0.56, p < 0.01). Trait divergence ( Q ST $$ {Q}_{\mathrm{ST}} $$ = 0.41–0.62) greatly exceeded neutral expectations ( F ST $$ {F}_{\mathrm{ST}} $$ = 0.0156), indicating diversifying selection on standing genetic variation. Genomic analyses identified PTM as the top GWAS candidate, a single candidate gene associated with flowering phenology variation, ranking in the top 5% of the genome-wide F ST $$ {F}_{\mathrm{ST}} $$ distribution and showing signatures of recent selective sweeps, with significantly higher expression in northern peripheral populations. In contras