SNP-slice resolves mixed infections: simultaneously unveiling strain haplotypes and linking them to hosts

Area of Science:

• Genomics
• Molecular Epidemiology
• Bioinformatics

Background:

• Multi-strain pathogen infections are common but challenging to analyze.
• Current methods often discard mixed infection samples, limiting understanding of pathogen genetic diversity and co-infection patterns.
• There is a need for scalable tools to analyze single nucleotide polymorphism (SNP) data from polygenomic samples.

Purpose of the Study:

• To develop a novel computational tool for resolving SNP-haplotypes and infection patterns from polygenomic pathogen data.
• To address the limitations of existing methods in analyzing mixed infection samples.
• To enhance the study of pathogen genetic diversity, co-infection dynamics, and genomic relatedness.

Main Methods:

• Development of a slice sampling Markov Chain Monte Carlo algorithm named SNP-Slice.
• The algorithm learns SNP-haplotypes and identifies infecting strains without requiring reference panels.
• Evaluation of SNP-Slice performance on empirical malaria and HIV datasets.

Main Results:

• SNP-Slice accurately reconstructs SNP-haplotypes and individual heterozygosities.
• The method outperforms state-of-the-art approaches in estimating infection multiplicity and allele frequencies.
• Demonstrated utility on real-world datasets, providing a new approach for molecular surveillance.

Conclusions:

• SNP-Slice offers a robust solution for analyzing polygenomic data in molecular epidemiology.
• The tool facilitates a deeper understanding of pathogen genetic diversity and co-infection dynamics.
• SNP-Slice opens new avenues for complex infection pattern resolution in pathogen surveillance.