TiTUS

TiTUS infers pathogen transmission trees from timed pathogen phylogenies combined with host-specific epidemiological data to reconstruct transmission histories in multi-strain infections and contexts with within-host diversity.

Key Features:

• Integration of Genomic and Epidemiological Data: Combines a timed pathogen phylogeny with host-specific epidemiological information such as entry-removal times and contact maps to constrain transmission histories.
• Direct Transmission Inference (DTI) Problem Formulation: Formulates the DTI problem to explicitly account for multi-strain infections and within-host diversity.
• SATISFIABILITY-Based Sampling: Uses SATISFIABILITY techniques to uniformly sample the space of feasible interval vertex labelings of the phylogeny while enforcing direct transmission constraints and supporting weak transmission bottlenecks.
• Parsimonious Transmission Trees: Prioritizes parsimonious transmission trees and summarizes sampled solutions using a consensus tree approach.
• Application to Simulated and Real Data: Validated on simulated datasets and documented real-world cases, including reconstruction of an HIV transmission chain.

Scientific Applications:

• Epidemiological reconstruction: Reconstructs transmission dynamics to study outbreak progression and inform analyses of outbreak structure.
• Multi-strain and within-host diversity studies: Applicable to pathogens exhibiting multi-strain infections and significant within-host diversity, such as HIV and influenza.
• Comparative evaluation of transmission scenarios: Enables exploration and comparison of alternative transmission histories consistent with genomic and epidemiological evidence.

Methodology:

Formulates the DTI problem using timed phylogenies and host epidemiological constraints (entry-removal times, contact maps), establishes hardness results for decision and counting versions, applies SATISFIABILITY techniques to uniformly sample feasible interval vertex labelings under direct transmission and weak bottleneck constraints, and selects parsimonious trees summarized via a consensus tree method.