Next Generation Malaria Imaging: AI-guided Analysis of Rhoptry Dynamics and Genome Remodelling in Malaria Parasite, Plasmodium falciparum
Sonja Frölich · 2025
Malaria, infecting over 200 million people and claiming 600,000 lives annually, relies on a critical 48-hour blood stage life cycle where Plasmodium falciparum invade red blood cells (RBCs). Interactions between the RBC surface and invasion ligands from merozoite rhoptries are essential for RBC invasion and parasitophorous vacuole (PV) formation. Within this PV, the parasite replicates via “closed” mitosis and unique cytokinesis, culminating in a multinucleated schizont, with each progeny merozoite inheriting the cellular machinery to function independently. To accelerate studies on parasite development within RBCs, we developed a machine learning (ML)-driven image analysis pipeline. Leveraging random forest algorithms, we automated the detection of parasite nuclei and rhoptries in super-resolved 3D volumes of immune-labelled infected RBCs. Our ML approach excels in accurately identifying nuclei in various replication stages and quantifying rhoptry structure, providing further insights into schizont maturation. The pipeline was trained on images of parasites at different developmental stages, enabling automated differentiation between pre- and post-replicating nuclei and structural rhoptry features associated with merozoite maturation. This novel ML-based method outperformed traditional techniques in speed and accuracy, particularly for proteins PfCERLI1, PfRAP1 and PfRON4, identifying additional organelles that were missed by traditional approaches. This ML-driven pipeline enhances analysis, enabling deeper exploration of nuclear replication and rhoptry function. It surpasses traditional methods and has potential for broader parasite research, demonstrating the value of ML in malaria research, expediting functional studies and drug discovery.