Improving statistical power in severe malaria genetic association studies by augmenting phenotypic precision

Area of Science:

• Genetics
• Epidemiology
• Malariology

Background:

• Severe falciparum malaria significantly impacts human evolution and genetic susceptibility studies.
• Phenotypic imprecision in diagnosing severe malaria, especially differentiating it from bacterial sepsis in young children, compromises genetic association findings.

Purpose of the Study:

• To develop and validate a probabilistic diagnostic model for severe malaria using readily available blood count data.
• To re-analyze existing genetic association data with improved case definition to reduce false discovery rates and increase statistical power.

Main Methods:

• A probabilistic diagnostic model was created using platelet and white blood cell counts.
• Clinical and genetic data from 2220 Kenyan children with clinically defined severe malaria and 3940 controls were re-analyzed.
• A data-tilting approach was proposed for case-control studies with potential phenotype mislabeling.

Main Results:

• The probabilistic model estimated that approximately one-third of clinically diagnosed severe malaria cases were misclassified.
• The model's performance was validated using the distribution of sickle cell trait.
• The proposed data-tilting approach reduced false discovery rates and improved statistical power in genome-wide association studies.

Conclusions:

• Accurate phenotyping is crucial for reliable genetic association studies of severe malaria.
• The developed probabilistic model offers a more precise method for diagnosing severe malaria in resource-limited settings.
• Improved diagnostic accuracy enhances the understanding of genetic susceptibility to severe malaria and its evolutionary impact.