vivaldi

vivaldi (Viral Variant Location and Diversity) is an R package for analyzing intrahost viral variation from Illumina sequencing data. The package is built around variant call format (VCF) files and provides tools to:

• import and tidy viral variant calls from one or many VCF files
• merge technical replicates
• filter variants by coverage and allele frequency thresholds
• parse SnpEff annotations
• summarize shared variants and mutation spectra
• quantify diversity using metrics such as Shannon entropy and dN/dS
• generate publication-ready visualizations of variant positions, frequencies, and genome-wide patterns

The package was developed for viral diversity analyses such as those described in the final published paper:

Roder AE, Johnson KEE, Knoll M, Khalfan M, Wang B, Schultz-Cherry S, Banakis S, Kreitman A, Mederos C, Wang W, Ruchnewitz D, Samanovic MI, Mulligan MJ, Lassig M, Łuksza M, Das S, Gresham D, Ghedin E. Optimized Quantification of Intrahost Viral Diversity in SARS-CoV-2 and Influenza Virus Sequence Data. mBio. 2023. https://doi.org/10.1128/mbio.01046-23

Installation

Install from CRAN

install.packages("vivaldi")

Install the development version from GitHub

# install.packages("remotes")
remotes::install_github("GreshamLab/vivaldi")

Package dependencies

vivaldi depends on widely used R packages for variant parsing, data wrangling, and visualization:

• vcfR for reading VCF files
• dplyr, tidyr, and magrittr for data manipulation
• ggplot2 and plotly for plotting
• seqinr for reading reference FASTA files
• glue for string handling

Input data

The main input is a directory of .vcf files. Depending on the workflow, you may also provide:

• a reference FASTA file used for alignment
• a table of segment or genome sizes
• a replicate mapping table for technical replicate merging
• SnpEff-annotated VCF files for annotation-aware downstream analyses

The package ships with example data accessible via system.file("extdata", package = "vivaldi"), including:

• H1N1.fa - example influenza reference FASTA
• SegmentSize.csv - segment size metadata
• reps.csv - replicate metadata
• vcfs/ - example annotated VCF files

An example processed dataset is also included as example_filtered_SNV_df.

Typical workflow

A common vivaldi analysis looks like this:

1. Load VCF files with arrange_data()
2. Merge technical replicates with merge_replicates() when replicate sequencing is available
3. Filter low-confidence variants with filter_variants()
4. Expand annotation fields with prepare_annotations()
5. Add metadata such as segment sizes with add_metadata()
6. Summarize diversity with functions such as tstv_ratio(), shannon_entropy(), and dNdS_segment()
7. Visualize results with functions such as af_distribution(), snv_location(), snv_genome(), snv_segment(), plot_shannon(), and shared_snv_plot()

Quick start

library(vivaldi)

vardir <- system.file("extdata", "vcfs", package = "vivaldi")
reference_fasta <- system.file("extdata", "H1N1.fa", package = "vivaldi")
seg_sizes <- system.file("extdata", "SegmentSize.csv", package = "vivaldi")
rep_info <- system.file("extdata", "reps.csv", package = "vivaldi")

vcf_df <- arrange_data(vardir, reference_fasta, annotated = "yes")
replicates <- read.csv(rep_info)
sizes <- read.csv(seg_sizes)

merged_df <- merge_replicates(
  vcf_df,
  replicates,
  "rep1",
  "rep2",
  c("sample", "CHROM", "POS", "REF", "ALT", "ANN", "ALT_TYPE", "major", "minor")
)

filtered_df <- filter_variants(
  merged_df,
  coverage_cutoff = 0,
  frequency_cutoff = 0.01
)

annotated_df <- prepare_annotations(filtered_df)
annotated_df <- add_metadata(annotated_df, sizes, c("CHROM"), c("segment"))
annotated_df <- shannon_entropy(annotated_df, genome_size = 13133)

plot_shannon(annotated_df)

Core functions

Import and preparation

• arrange_data() - read VCF files and combine them into a single dataframe
• read_reference_fasta_dna() - extract chromosome or segment sizes from a FASTA file
• prepare_annotations() - split SnpEff annotations into separate columns
• snpeff_info() - parse annotation information from VCF INFO fields
• add_metadata() - join external metadata to the variant dataframe

Filtering and replicate handling

• filter_variants() - apply coverage and allele frequency thresholds
• merge_replicates() - keep shared variants across sequencing replicates and compute summary frequencies

Summary statistics

• tally_it() - count variants over user-defined groups
• tstv_ratio() - calculate transition/transversion ratios
• shannon_entropy() - calculate per-position, per-segment, and genome-wide diversity metrics
• dNdS_segment() - estimate dN/dS summaries by coding feature or segment

Visualization and exploration

• af_distribution() - plot minor allele frequency distributions
• position_allele_freq() - inspect the allele frequencies of a single site across samples
• shared_snv_plot() and shared_snv_table() - identify and summarize shared variants
• snv_location() - visualize SNV positions across samples and segments
• snv_genome() - summarize SNVs across genomes
• snv_segment() - summarize SNVs by genome segment
• plot_shannon() - visualize Shannon entropy summaries
• tstv_plot() - visualize transition/transversion summaries

Documentation

For a longer worked example, see the package vignette:

browseVignettes("vivaldi")

Source documentation and examples are also available in the package help pages:

help(package = "vivaldi")

Testing and development

The repository uses standard R package tooling, including testthat tests under tests/testthat and GitHub Actions R-CMD-check workflows.

Authors

Marissa Knoll, Katherine Johnson, Megan Hockman, Eric Borenstein, Mohammed Khalfan, Elodie Ghedin, and David Gresham

Maintainer: David Gresham dg107@nyu.edu

Citation

If you use vivaldi in published research, please cite the published article rather than the preprint:

Roder AE, Johnson KEE, Knoll M, Khalfan M, Wang B, Schultz-Cherry S, Banakis S, Kreitman A, Mederos C, Wang W, Ruchnewitz D, Samanovic MI, Mulligan MJ, Lassig M, Łuksza M, Das S, Gresham D, Ghedin E. Optimized Quantification of Intrahost Viral Diversity in SARS-CoV-2 and Influenza Virus Sequence Data. mBio. 2023. https://doi.org/10.1128/mbio.01046-23