Comparative Genomics and Network Modeling of Parasites

Eukaryotic parasites cause diseases with large global burdens, resulting in over one million deaths annually. To develop novel antiparasitic drugs, experimental model systems often use alternative parasite species that are easier to manipulate genetically or to observe in vivo in mice, extrapolating these observations to disease-causing parasites. However, characterization of functional differences between parasite species is limited to post hoc and single-target studies, limiting the utility of these experimental models of disease and historic data. Each parasite genome encodes unique enzyme annotations; however, it is unclear whether these differences arise from divergent functions or incomplete genome annotation.

To address this challenge, we generated metabolic reconstructions from 160 parasite genomes, representing 38 genera and 111 species. Unsurprisingly, one of the largest genomes (Chromera velia) has the most unique reactions (35); however, the smallest genome (Plasmodium billcollinsi) has six unique reactions. Reconstruction and genome sizes are correlated, but even small networks contain unique features. We next applied our novel automated curation approach to leverage manual curation to improve reconstructions for related organisms. We used these semi-curated reconstructions to compare metabolic capacity and pathway utilization and generate hypotheses about species-specific functions. Here, we focus on the most lethal malaria parasite, Plasmodium falciparum, and the mouse model of severe malaria, P. berghei. For example, while P. falciparum grows in anucleated host erythrocytes, berghei prefers nucleated reticulocytes; in silico, P. falciparum has a reduced demand for host purines when compared to P. berghei, perhaps explaining this host cell preference and highlighting a pathway for which P. berghei is a poor model organism for P. falciparum antimalarial development. By performing these analyses with all 160 reconstructions, we developed a framework to identify metabolic discrepancies and commonalities between genera and species, facilitating comparison of experimental findings and optimizing model system selection to develop therapeutics for parasitic diseases.