K-Mer Analyses Reveal Different Evolutionary Histories of Alpha, Beta, and Gamma Papillomaviruses

Zigui Chen, Filippo Utro, Daniel E. Platt, Rob DeSalle, Laxmi Parida, Paul K.s. Chan, Robert D. BurkInternational Journal of Molecular Sciences 2021 · DOI

TL;DR

This study uses k-mer analysis, a sequence alignment-free method, to explore the evolutionary history of human papillomaviruses (HPVs). The findings reveal distinct evolutionary paths for Alpha, Beta, and Gamma HPV genera, suggesting that these viruses have undergone different selective pressures and host adaptations. This work highlights the power of orthogonal approaches in uncovering novel insights into viral genome evolution.

What Problem Does This Paper Address?

Papillomaviruses (PVs) are a diverse group of DNA viruses that infect a wide range of species, with human papillomaviruses (HPVs) being particularly well-known for their association with various cancers. HPVs are classified into several genera, with Alpha, Beta, and Gamma PVs representing the majority. While traditional phylogenetic methods based on sequence alignment have been used to study their evolution, the complex evolutionary histories of PVs, including potential recombination events and rapid divergence, can obscure clear phylogenetic signals. This makes it challenging to fully understand the evolutionary features that drive their divergence, tissue tropism, and pathogenicity. This paper aims to address this by employing alignment-free methods to uncover hidden biological signals and provide a more comprehensive understanding of PV evolution.

Methods

The researchers utilized a combination of approaches:

  • K-mer Distribution Analysis: This is the core alignment-free method employed. They analyzed the frequency of "trimers" (k=3 nucleotide sequences) across the entire PV genomes. This approach captures compositional biases and patterns without relying on direct sequence alignment, making it robust to insertions, deletions, and recombination. A hierarchical clustering based on Kullback-Leibler distances of these trimer distributions was used to construct a phylogeny.
  • Alignment-Based Phylogenetic Analysis: For comparison, they also performed traditional phylogenetic analysis using maximum likelihood (ML) trees based on concatenated nucleotide sequences of conserved ORFs (E1, E2, L1, L2).
  • Genetic Metrics: They calculated various genetic metrics, including:
  • Effective Number of Codons (ENC): To assess codon usage bias.
  • Relative Synonymous Codon Usage (RSCU): To quantify the preference for specific codons.
  • Dinucleotide Suppression Analysis: To measure the observed vs. expected frequencies of dinucleotides (e.g., CpG, ApT), which can reflect mutational pressures and host defense mechanisms.
  • Correspondence Analysis (RDA): Used to explore correlations between codon usage bias and genomic composition.

Datasets

The study analyzed a comprehensive dataset of 640 PV genomes, including:

  • 442 human PV (HPV) types.
  • 27 non-human primate PV (NHP-PV) types.
  • 171 non-primate animal PV (NPA-PV) types.

A refined dataset of 228 PV genomes was used for detailed analyses of codon usage bias and dinucleotide suppression, comprising 214 human PV complete genomes and 14 macaque PV types. These sequences were retrieved from public databases such as the HPV Center, Papillomavirus Episteme (PAVE), and GenBank/NCBI.

Key Insights and Ideas

  • Gamma-PVs are the Most Diverse: The study found that Gamma-PVs exhibit the highest genetic diversity among the analyzed genera, suggesting they may have experienced less stringent selective pressures compared to Alpha and Beta PVs.
  • Orthogonal Approaches Reveal Hidden Relationships: The alignment-free k-mer analysis revealed phylogenetic relationships that differed from traditional alignment-based methods. This highlights the value of using multiple, complementary analytical tools to understand viral evolution.
  • Low-Risk Alpha HPVs Have a Unique Evolutionary Trajectory: A subset of low-risk Alpha HPVs (species Alpha-2, 3, 4, and 14) clustered with non-human primate PVs in the k-mer phylogeny, suggesting a distinct evolutionary history and potential host specificity.
  • Gamma-6 HPV Types Cluster with Alpha HPVs: Surprisingly, Gamma-6 species types, typically found in the cervicovaginal region, were grouped with the main Alpha-HPV clade in the trimer-constructed phylogeny. This suggests a shared evolutionary history or functional convergence that wasn't apparent from homology-based methods.
  • Codon Usage and Dinucleotide Patterns Reflect Host Specificity and Tropism: The analysis of codon usage bias and dinucleotide suppression revealed patterns that correlate with host specificity (human vs. macaque PVs) and tissue tropism (mucosal vs. cutaneous HPVs). For instance, GpA dinucleotide suppression was more pronounced in Alpha-HPVs, potentially linked to host immune responses.

Key Results

  • The k-mer (trimer) based phylogeny largely recapitulated the major HPV genera (Alpha, Beta, Gamma) but also revealed distinct groupings, such as the separation of Alpha-HPV-LR2 from other HPVs and its clustering with non-human primate PVs.
  • Gamma-PVs showed significantly higher inter-species genetic diversity (41.9 ± 1.8%) compared to Alpha-PVs (39.7 ± 2.1%) and Beta-PVs (35.9 ± 1.4%).
  • Codon usage analysis indicated that Alpha-HPV-LR2 had significantly higher ENC values compared to other HPVs, suggesting a less biased codon usage.
  • Dinucleotide analysis revealed significant differences in suppression patterns between Alpha-HPVs and Beta-/Gamma-HPVs, particularly for GpA (=TpC), which was more depleted in Alpha-HPVs.
  • A strong correlation was observed between the loss of CpG dinucleotides and the gain of non-CpG methylation dinucleotides (ApG, CpA, CpC) in certain HPV groups, suggesting a role for host methylation in shaping viral genomes.

Limitations and Future Work

  • Limitations: While the study used a large dataset, the authors acknowledge that the "imperfect boundary between clinical manifestation and phylogenetic relationships for HPVs may imply complex evolutionary histories." The interpretation of k-mer distributions can be complex, and while it offers an alignment-free approach, it might not capture all nuances of sequence evolution. The study focused on a subset of PV genera, and further investigation into other genera like Mu- and Nu-PV would be beneficial.
  • Future Work: The authors suggest that characterizing papillomavirus heterogeneity via orthogonal approaches reveals novel insights into the biological understanding of HPV genomes. Future work could involve expanding these analyses to include a wider range of PV genera and exploring the functional implications of the observed evolutionary patterns in more detail, particularly concerning host adaptation, immune evasion, and pathogenicity. The study also points to the need for further investigation into the dynamics and role of CpG methylation in HPV genomes and its impact on viral replication and disease progression.

Summary updated

Abstract

Papillomaviruses (PVs) are a heterogeneous group of DNA viruses that can infect fish, birds, reptiles, and mammals. PVs infecting humans (HPVs) phylogenetically cluster into five genera (Alpha-, Beta-, Gamma-, Mu- and Nu-PV), with differences in tissue tropism and carcinogenicity. The evolutionary features associated with the divergence of <i>Papillomaviridae</i> are not well understood. Using a combination of <i>k</i>-mer distributions, genetic metrics, and phylogenetic algorithms, we sought to evaluate the characteristics and differences of Alpha-, Beta- and Gamma-PVs constituting the majority of HPV genomes. A total of 640 PVs including 442 HPV types, 27 non-human primate PV types, and 171 non-primate animal PV types were evaluated. Our analyses revealed the highest genetic diversity amongst Gamma-PVs compared to the Alpha and Beta PVs, suggesting reduced selective pressures on Gamma-PVs. Using a sequence alignment-free trimer (<i>k</i> = 3) phylogeny algorithm, we reconstructed a phylogeny that grouped most HPV types into a monophyletic clade that was further split into three branches similar to alignment-based classifications. Interestingly, a subset of low-risk Alpha HPVs (the species Alpha-2, 3, 4, and 14) split from other HPVs and were clustered with non-human primate PVs. Surprisingly, the trimer-constructed phylogeny grouped the Gamma-6 species types originally isolated from the cervicovaginal region with the main Alpha-HPV clade. These data indicate that characterization of papillomavirus heterogeneity via orthogonal approaches reveals novel insights into the biological understanding of HPV genomes.
BibTeX
@article{chen2021k,
  author = {Zigui Chen and Filippo Utro and Daniel E. Platt and Rob DeSalle and Laxmi Parida and Paul K.S. Chan and Robert D. Burk},
  title = {{K-Mer Analyses Reveal Different Evolutionary Histories of Alpha, Beta, and Gamma Papillomaviruses}},
  journal = {International Journal of Molecular Sciences},
  year = {2021},
  doi = {10.3390/ijms22179657},
}
Figure