Abstract
Molecular diagnostic approaches are increasingly included in the diagnostic workup and even in the primary diagnosis of malaria in non-endemic settings, where it is difficult to maintain skillful microscopic malaria detection due to the rarity of the disease. Pathogen-specific nucleic acid amplification, however, bears the risk of overlooking other pathogens associated with febrile illness in returnees from the tropics. Here, we assessed the discriminatory potential of metagenomic sequencing for the identification of different Plasmodium species with various parasitemia in EDTA blood of malaria patients. Overall, the proportion of Plasmodium spp.-specific sequence reads in the assessed samples showed a robust positive correlation with parasitemia (Spearman r = 0.7307, p = 0.0001) and a robust negative correlation with cycle threshold (Ct) values of genus-specific real-time PCR (Spearman r = -0.8626, p ≤ 0.0001). Depending on the applied bioinformatic algorithm, discrimination on species level was successful in 50% (11/22) to 63.6% (14/22) instances. Limiting factors for the discrimination on species level were very low parasitemia, species-depending lacking availability of reliable reference genomes, and mixed infections with high variance of the proportion of the infecting species. In summary, metagenomic sequencing as performed in this study is suitable for the detection of malaria in human blood samples, but the diagnostic detection limit for a reliable discrimination on species level remains higher than for competing diagnostic approaches like microscopy and PCR.