Improving Malaria Parasite Gene Flow Inference Under Sparse Spatial Sampling

Area of Science:

• Population genetics
• Genomic epidemiology
• Parasitology

Background:

• Estimating malaria parasite migration is vital for targeted elimination strategies.
• Existing methods like EEMS can produce misleading gene flow maps with sparse or uneven sampling.
• Posterior probabilities alone do not guarantee geographic reliability of inferred migration patterns.

Purpose of the Study:

• To develop and validate a sample location-aware (SLA) filtering workflow for georeferenced gene flow data.
• To improve the accuracy and interpretability of spatial parasite migration patterns.
• To address artefacts in gene flow mapping caused by inadequate geographic sampling.

Main Methods:

• Developed an SLA filtering workflow using topological skeletons and kernel density estimation.
• Applied the workflow to assess sample density around inferred migration features.
• Validated the method using simulated genomic data and an empirical Plasmodium falciparum dataset from Cambodia.

Main Results:

• SLA filtering significantly improves the precision of inferred low-migration regions compared to posterior probability filtering alone.
• Posterior-only filtering can retain spurious migration barrier features in sparsely sampled areas.
• SLA filtering yields more stable parasite migration estimates under location subsampling.

Conclusions:

• Incorporating geographic sampling density enhances the reliability of spatial gene flow analyses.
• SLA filtering offers a generalizable diagnostic and filtering approach for population-genetic migration mapping.
• This method improves the interpretability of malaria parasite migration patterns for elimination efforts.