Transmission lineage dynamics and the detection of viral importation in emerging epidemics

The accurate inference of pathogen movements among locations during an epidemic is crucial for measuring infectious disease spread and for informing effective control strategies. Phylogeographic methods can reconstruct historical patterns of disease dissemination by combining the evolutionary history of sampled pathogen genomes with geographic information. Despite a substantial expansion of pathogen genomics during and after the COVID-19 pandemic, only a small fraction of infections are typically sampled and sequenced, leading to underestimation of the true intensity of viral importation. Here, we seek to understand the sampling processes underlying this underestimation. We show that the coupling of viral importation and local transmission dynamics can result in local transmission lineages with different size distributions, influencing the probability that individual viral importation events will be detected. Using analytical and simulation approaches, we show that both the proportion of importation events detected and the temporal patterns of inferred importation are highly sensitive to importation dynamics and local transmission parameters, resulting in substantial biases, particularly under low-intensity sampling. Our findings highlight the importance of interpreting phylogeographic estimates in the context of outbreak conditions, particularly when comparing viral movements across time and among epidemic settings characterised by rapid spatial dissemination. These insights are critical for improving the reliability of genomic epidemiology approaches to the design of public health responses.