The authors constructed a multi-scale within-host phylodynamic model named WiPhy to investigate how viral replication, immune responses, and mutation together shape intrahost diversity and the risk of transmitting novel variants. The model couples processes of viral growth and decline with both innate immune responses and acquired immune responses, while allowing accumulation of mutations and selection of variants during an individual's infection.
WiPhy is presented as a tool to link mechanistic within-host biology to measurable phylodynamic outcomes and transmission risk. The abstract indicates the model was used to generate predictions about when and how transmissible variants are likely to emerge during the course of infection.
The authors report that they validated the WiPhy model against quantitative viral and phylodynamic metrics. Specific validation metrics, datasets, or statistical results are not detailed in the abstract. The source indicates additional resources (code repository) are provided, but the abstract itself does not enumerate the precise validation procedures or numerical fit statistics.
According to the model output, most acute infections rapidly generate considerable genetic diversity through the accumulation of minor variants. However, in the majority of such infections these minor variants do not reach concentrations sufficient for onward transmission. Thus, while intrahost diversity is common, it usually does not translate into between-host evolutionary change because of low variant frequencies and transmission bottlenecks.
This pattern underscores a distinction between frequent within-host mutation and comparatively rare between-host establishment of new, fitter lineages.
The model predicts that a delayed innate immune response correlates with higher peak viral load and greater intrahost diversification. When innate responses are slow to act, viral replication continues unchecked for longer periods, producing more genetic variants and raising the chance that either the founder virus or a novel variant is transmitted to another host.
However, the abstract notes that variants transmitted under these conditions are often equally or less fit than the founder virus; delayed innate responses raise transmission risk primarily by increasing viral quantity and diversity rather than by reliably selecting fitter mutations.
The risk of transmitting a fitter variant — defined in the model as one bearing fitness-enhancing non-synonymous mutations — is concentrated in a minority of infections. Specifically, the model identifies that approximately ~10% of infections where viral loads remain sufficiently high after 10–14 days present the greatest opportunity for selection and transmission of fitter variants.
In these prolonged infections, either a non-sustained innate response or a weak acquired immune response (or both) allow enough time for selection to act on new non-synonymous mutations, increasing the probability that a variant with enhanced replicative or immune-evasive fitness will rise to transmitable frequencies.
Using a simulated cohort of roughly 1,500 individuals, the model produced population-level estimates of transmission risk attributable to variants with enhanced fitness. Across that cohort, 5% of transmission risk was attributed to variants that had acquired enhanced fitness through nonsynonymous mutations. Additionally, transmission risk from fitter variants was highly skewed: 13% of simulated infections accounted for 90% of the fitter-variant transmission risk. These results indicate that a small subset of infections — the prolonged, high-viral-load cases — disproportionately contributes to the emergence and spread of fitter variants.
The abstract does not provide further breakdowns, confidence intervals, or sensitivity analyses for these percentages; methodological details and supplementary results are referenced but not described in the abstract itself.
The authors conclude that the timing and interaction of viral replication and host immune responses within individuals generate strong bottlenecks that largely restrict between-host evolutionary change. Although intrahost diversity is routinely produced, most variants fail to reach transmissible frequencies.
A critical implication is that interventions or circumstances that reduce the frequency or duration of prolonged, high-viral-load infections — for example, interventions that speed innate responses or strengthen acquired immunity — could reduce opportunities for selection of fitter variants. The abstract does not prescribe specific clinical interventions or public health measures, and does not report empirical trials testing such strategies; it presents model-derived insights.
All points above are drawn from the abstract of the preprint. The abstract reports the model name (WiPhy), core predictions, and numerical summaries for simulated cohorts, but it does not provide detailed methods, model parameters, validation datasets, or full sensitivity analyses in the abstract text. Those details and supplementary materials are indicated as available in the source (including a linked repository), but they were not reported within the abstract itself.
Readers should note the article is a preprint and has not undergone peer review, as stated by the authors and the hosting repository.