Using earth mover's distance for viral outbreak investigations

Area of Science:

• Virology
• Genomics
• Epidemiology

Background:

• RNA viruses exhibit high mutation rates, enabling immune evasion and posing public health threats.
• Accurate inference of transmission clusters is critical for managing viral outbreaks.
• Next-generation sequencing (NGS) offers powerful tools for outbreak analysis, but current methods rely on time-consuming sequence assembly.

Purpose of the Study:

• To develop and validate an algorithm for inferring viral transmission characteristics directly from raw NGS reads.
• To bypass the need for complex and error-prone viral population reconstruction.

Main Methods:

• The study proposes a novel algorithm that analyzes raw NGS reads to identify genetic relatedness between viral samples.
• The method avoids the traditional requirement of assembling short reads into complete sequences.

Main Results:

• Experimental validation using Hepatitis C Virus (HCV) outbreak data demonstrated the algorithm's success.
• The algorithm accurately identified genetic relatedness, inferred transmission direction and clusters, and pinpointed outbreak sources.
• It successfully determined the presence and identity of outbreak sources within sequenced samples.

Conclusions:

• The developed algorithm effectively clusters genetically related viral samples and predicts outbreak sources.
• It reconstructs key transmission dynamics, including directionality and clustering.
• A significant advantage is the ability to use raw NGS reads, eliminating assembly-related errors and reducing processing time.