Abstract
SARS-CoV-2, the virus responsible for the COVID-19 pandemic, evolves into different genetic variants by accumulating mutations as it spreads globally. Beyond the diversity of consensus genomes across patients, RNA viruses can also display significant genetic diversity within individual hosts. This study systematically examines the intra-patient genetic diversity of SARS-CoV-2 using a large cohort of 3,939 publicly available deeply sequenced genomes and 749 samples from Switzerland. We found that the distribution of diversity across patients and genomic loci is highly unbalanced, with a minority of hosts and positions accounting for most of the variation. Notably, the D614G variant in the Spike gene exhibits high intra-host diversity in nearly 30% of the public cohort. Furthermore, patient age was identified as a significant predictor of viral genetic diversity, suggesting that older individuals may harbor more diverse viral populations.
Introduction: The Viral Swarm
Most genomic studies of SARS-CoV-2 focus on the "consensus" sequence—the dominant version of the virus found in a patient. However, RNA viruses like SARS-CoV-2 are notorious for their high mutation rates, often existing within a single host as a collection of closely related variants known as a quasispecies. Understanding this intra-patient diversity is critical because co-existing variants can influence disease progression, transmission dynamics, and the potential for immune evasion or resistance to medical treatments. This research provides a large-scale survey of this "internal" diversity across thousands of patients.
Methods and Pipeline
The researchers utilized a specialized bioinformatics pipeline called V-pipe to process raw deep-sequencing data. Unlike standard tools that identify only the dominant sequence, V-pipe is designed to call low-frequency single-nucleotide variants (SNVs)—mutations appearing in only a small percentage of the viral population within a patient.
To quantify diversity, Shannon entropy was calculated for every position in the viral genome. This metric allows researchers to distinguish between highly conserved sites (low entropy) and "hotspots" where multiple variants coexist (high entropy). Regression modeling was then used to determine if epidemiological factors like age or sex could predict viral mutation levels, while carefully controlling for technical noise such as sequencing depth and error rates.
Datasets
The study analyzed two major cohorts:
- Public Cohort: 3,939 deeply sequenced genomes retrieved from the Sequence Read Archive (SRA), primarily using Illumina technology.
- Swiss Cohort: 749 recently sequenced samples from Switzerland, serving as an independent validation set.
Key Insights: Hotspots and the Age Factor
The analysis revealed several critical patterns in how SARS-CoV-2 diversifies within its host:
- The "Heavy Tail" Distribution: Viral diversity is not spread evenly. A small minority of patients and specific genomic locations account for the vast majority of observed genetic variation.
- The Age Factor: The most striking finding is that viral genetic diversity increases with the age of the patient. The model predicts an 8.6% increase in total entropy for every decade of age, suggesting that in older patients, the virus has more opportunity to diversify, possibly due to a slower or less effective immune response.
- D614G Coexistence: The famous D614G mutation in the Spike protein, known for increasing infectivity, is not always fixed. In nearly 30% of the public cohort, the researchers found both the original and mutated variants co-existing within the same individual.
Key Results
- Age as a Predictor: Age was a significant predictor of diversity ($p = 4 \times 10^{-4}$), a finding corroborated in the Swiss cohort using sequencing date as a proxy for age.
- Gene-Specific Diversity: The Matrix (M) gene was found to be the most diverse on average, while ORF7b was the most stable.
- Mutation Frequencies: At the most diverse site (position 11075), nearly half of the samples showed some level of mutation, frequently involving low-frequency deletions that could alter the resulting protein.
- Technical Robustness: While sequencing technology affects detection sensitivity, the biological signal regarding age remained consistent across different platforms.
Limitations and Future Outlook
Detecting subclonal mutations is highly sensitive to sequencing depth and potential artifacts from PCR amplification. While the researchers adjusted for these technical parameters, "noise" remains an inherent challenge in deep-sequencing studies.
A major remaining question is causality: Does higher viral diversity drive the worse clinical outcomes seen in the elderly, or is the diversity simply a result of a less effective immune system allowing the virus to replicate and mutate longer? Future research integrating high-quality clinical data with these genetic profiles will be essential to determine if "internal" viral evolution can be used to predict disease severity or transmission risk.
Summary updated