Phylogenetic meta-analysis of chronic SARS-CoV-2 infections in immunocompromised patients shows no evidence of elevated evolutionary rates

Abstract

Genomic sequences from rapidly evolving pathogens, sampled over time, hold information on disease origin, transmission, and evolution. Together with their sampling times, sequences can be used to estimate the rates of molecular evolution and date evolutionary events through molecular tip-dating. The validity of this approach, however, depends on whether detectable levels of genetic variation have accumulated over the given sampling interval, generating temporal signal. Moreover, different molecular dating methods have demonstrated varying degrees of systematic biases under different biologically realistic scenarios, such as the presence of phylo-temporal clustering. Chronic SARS-CoV-2 infection in immunocompromised patients has been linked to remarkably higher intra-host molecular rates than those of global lineages, facilitating the emergence of novel viral lineages. Yet, most studies reporting accelerated rates lack the evaluation of temporal signal or comparison of multiple methods of inference, both required to reliably estimate molecular rates. In this study, we use 26 previously published longitudinally sampled sequence series obtained from chronically infected immunocompromised patients to re-evaluate the rate of SARS-CoV-2 intrahost evolution. Using a range of methods, we analyse the strength of temporal signal and infer evolutionary rates from tip-calibrated phylogenies. Regardless of heterogeneity in rate estimates between sample series and methods, we find within-host rates to be in good agreement with rates derived from host-to-host transmission chains. Our findings suggest that when certain limitations of the methodology are disregarded, such as the underlying assumption of phylogenetic independence or the methods sensitivity to phylo- temporal grouping, evolutionary rates can be substantially overestimated. We demonstrate that estimating within-host rates is a challenging question necessitating careful interpretation of findings. While our results do not support faster evolution across the complete viral genome during chronic SARS-CoV-2 infection, prolonged viral shedding together with relapsing viral load dynamics may nevertheless promote the emergence of new viral variants in immunocompromised patients. AUTHOR SUMMARYThe evolutionary origin of SARS-CoV-2 variants of concern (VOC) is a longstanding point of controversy, with multiple proposed explanations. Observations of immunocompromised individuals being at a greater risk of developing a prolonged SARS-CoV-2 infection have led to the Chronic infection hypothesis, suggesting that these cases may contribute to the emergence of VOCs. Correspondingly, many studies have reported accelerated viral evolution of SARS-CoV-2 within immunocompromised individuals with respect to the viral background population. However, many of these findings have not been validated with appropriate analytical methods. In this study we re-evaluate the rate of intrahost viral evolution of SARS- CoV-2 within immunocompromised patients utilising a range of methods. We assess the performance of different methodologies and compare our results to published estimates of SARS-CoV-2 evolutionary rates. Our systematic comparison showed no evidence supporting the previous claims of elevated levels of intrahost evolution in immunocompromised patients with chronic SARS-CoV-2. Instead, our findings exemplify the complexity of within-host viral dynamics, suggesting that a more comprehensive understanding of SARS-CoV-2 evolutionary processes would be derived from concurrent evaluation of viral genomic data together with patients clinical information.