Widespread PCR-generated chimeras confound Plasmodium species identification and zoonotic malaria surveillance

The polymerase chain reaction (PCR) is widely used to amplify malaria parasite DNA to diagnose Plasmodium infections, but can generate artefactual chimeric sequences through template switching. Here, we examine the occurrence and consequences of PCR-generated chimeras in studies of malaria parasites from humans, macaques and mosquitoes, and highlight two major sources of errors. First, in samples containing mixed-species infections, PCR can produce interspecific chimeras. This problem is particularly acute among closely related species of Plasmodium that infect macaques in southeast Asia, several of which have been transmitted to humans. Reanalysis of published mitochondrial DNA sequences from Thailand and Malaysia revealed numerous chimeric sequences involving P. knowlesi, P. inui, and P. cynomolgi, as well as P. vivax and other related species. Second, amplification of multicopy genes can also generate intra-genomic chimeras. The most widely used PCR template for malaria parasite identification has been the small subunit (18S) ribosomal RNA gene, but Plasmodium genomes contain two divergent forms of the rRNA unit, the A-type and the S-type. We show that several studies have generated intertype rRNA chimeras, which can occur even in the absence of multi-species infection. This has led to difficulties in identifying the relevant species, given an inflated view of sequence diversity, and has confused conclusions regarding zoonotic transmissions. These findings highlight the need for more rigorous approaches to PCR amplification to avoid the generation of chimeras, and for the validation of potentially chimeric sequences before phylogenetic or epidemiological interpretation.