BACKGROUND:Current methods to classify local and imported malaria infections depend primarily on patient travel history, which can have limited accuracy. Genotyping has been investigated as a complementary approach to track the spread of malaria and identify the origin of imported infections. METHODS:An extended panel of 26 microsatellites (16 new microsatellites) for Plasmodium falciparum was evaluated in 602 imported infections from 26 sub-Saharan African countries to the Jiangsu Province of People's Republic of China. The potential of the 26 microsatellite markers to assign imported parasites to their geographic origin was assessed using a Bayesian method with Markov Chain Monte Carlo (MCMC) as implemented in the program Smoothed and Continuous Assignments (SCAT) with a modification to incorporate haploid genotype data. RESULTS:The newly designed microsatellites were polymorphic and are not in linkage disequilibrium with the existing microsatellites, supporting previous findings of high rate of recombination in sub-Saharan Africa. Consistent with epidemiology inferred from patients' travel history, no evidence for local transmission was found; nearly all genetically related infections were identified in people who travelled to the same country near the same time. The smoothing assignment method assigned imported cases to their likely geographic origin with an accuracy (Angola: 59%; Nigeria: 51%; Equatorial Guinea: 40%) higher than would be achieved at random, reaching statistical significance for Angola and Equatorial Guinea. CONCLUSIONS:Genotyping using an extended microsatellite panel is valuable for malaria case classification and programme evaluation in an elimination setting. A Bayesian method for assigning geographic origin of mammals based on genetic data was adapted for malaria and showed potential for identification of the origin of imported infections.