MLST_F

MLST_F deconvolutes within-host bacterial strain diversity from whole-genome sequencing (WGS) data using a multi-locus sequence typing (MLST) framework to infer allele sets, allele proportions, and strain types.

Key Features:

• MLST-based framework: Uses MLST derived from WGS data to characterize genomic variability of pathogens within a single host by locus-specific alleles.
• Two-stage processing methodology: Implements a two-stage approach where samples are first assigned allele sets and allele proportions for each MLST locus and then associated with strain types while minimizing the number of unobserved strains and maintaining genetic proximity to observed alleles.
• Optimization via Mixed Integer Linear Programming (MILP): Employs MILP to assign strain types that respect observed allele proportions and the optimization criteria.
• Application versatility: Demonstrated with Borrelia burgdorferi and applicable to any bacterial pathogen with an established MLST scheme.

Scientific Applications:

• Evolutionary studies: Capture within-host genomic diversity to investigate pathogen evolutionary adaptations.
• Host-pathogen interaction analysis: Resolve strain-level composition to study differential interactions of strains with host environments.
• Disease transmission patterns: Inform transmission dynamics and epidemiological analyses through accurate within-host strain typing.

Methodology:

Process WGS-derived MLST data in two stages: per-sample assignment of alleles and allele proportions for each MLST locus, followed by strain-type association via MILP optimization that minimizes novel strains while maintaining genetic closeness to observed alleles.