poodleR

poodleR helps estimate donor proportions in pooled sequencing experiments using:

• donor genotypes stored in a VCF
• pooled read counts exported by bam-readcount

The package covers the full workflow:

1. Convert donor genotypes to minor-allele dosages with create_gt_dosages().
2. Convert pooled read counts to minor-allele frequency estimates with estimate_b_from_bam_readcount().
3. Run the combined preprocessing step with calculate_nonRef_vector_and_genotype_df().
4. Estimate donor weights with estimate_weights().
5. Use is_identical_genotype() and mod_lsqlincon() as lower-level utilities.

Install from GitHub

install.packages("remotes")

remotes::install_github(
  "TrynkaLab/poodleR",
  dependencies = TRUE,
  build_vignettes = TRUE
)

Publishing checklist

Before announcing the package publicly:

1. Install the package dependencies locally.
2. Run devtools::check() or R CMD check.
3. Confirm the vignette builds and the example workflow runs.
4. Push the repository to a public GitHub repo named poodleR.
5. Update the USERNAME/poodleR placeholders in this README.
6. Create a release tag such as v0.1.0.

This repository now includes a GitHub Actions workflow at .github/workflows/R-CMD-check.yaml so every push and pull request can run package checks automatically.

Example files shipped with the package

The repository includes compressed example inputs in inst/extdata:

bam_path <- system.file("extdata", "bam-readcount_example.tsv.gz", package = "poodleR")
vcf_path <- system.file("extdata", "vcf_example.vcf.gz", package = "poodleR")

It also includes smaller .rda example objects in data/ for quick experimentation:

• A: genotype dosage table
• b: pooled allele-frequency table with b_estimate
• b_estimate_dt: filtered allele-frequency table aligned to A
• bam_readcount: imported bam-readcount example
• vcf: donor genotype VCF as a vcfR object

Quick start

library(poodleR)

bam_path <- system.file("extdata", "bam-readcount_example.tsv.gz", package = "poodleR")
vcf_path <- system.file("extdata", "vcf_example.vcf.gz", package = "poodleR")

prepared <- calculate_nonRef_vector_and_genotype_df(
  bam_readcount_path = bam_path,
  vcf_path = vcf_path,
  min_read_overlap = 5
)

weights <- estimate_weights(
  b = prepared$b_estimate_table$b_estimate,
  A = prepared$ma_dosages
)

weights
sum(weights)

Tutorial for all exported functions

create_gt_dosages()

Convert a vcfR object to a donor-by-SNP dosage table:

data(vcf, package = "poodleR")

dosages <- create_gt_dosages(vcf)
head(dosages[, 1:4])

Each donor column is numeric with values 0, 0.5, or 1, and rn stores the SNP identifier (CHROM_POS_REF).

estimate_b_from_bam_readcount()

Convert imported bam-readcount rows to pooled allele frequencies after aligning them to VCF alleles:

data(bam_readcount, package = "poodleR")
data(vcf, package = "poodleR")

colnames(bam_readcount) <- c("CHROM", "POS", "REF", "TOTAL_READS", "DEL", "A", "C", "G", "T", "N")
bam_readcount$CHROM_POS_REF <- paste(bam_readcount$CHROM, bam_readcount$POS, bam_readcount$REF, sep = "_")

vcf_reduced <- data.table::as.data.table(vcf@fix[, 1:5])
vcf_reduced$CHROM_POS_REF <- paste(vcf_reduced$CHROM, vcf_reduced$POS, vcf_reduced$REF, sep = "_")
vcf_reduced <- vcf_reduced[vcf_reduced$CHROM_POS_REF %in% bam_readcount$CHROM_POS_REF, ]
bam_readcount <- bam_readcount[bam_readcount$CHROM_POS_REF %in% vcf_reduced$CHROM_POS_REF, ]

b_estimate_table <- estimate_b_from_bam_readcount(bam_readcount, vcf_reduced)
head(b_estimate_table)

calculate_nonRef_vector_and_genotype_df()

Run the full preprocessing pipeline directly from the shipped example files:

prepared <- calculate_nonRef_vector_and_genotype_df(
  bam_readcount_path = bam_path,
  vcf_path = vcf_path,
  min_read_overlap = 5
)

names(prepared)
nrow(prepared$b_estimate_table)
nrow(prepared$ma_dosages)

This returns two aligned tables ready for deconvolution:

• b_estimate_table: pooled allele frequencies
• ma_dosages: donor genotype dosages

estimate_weights()

Estimate donor contributions from the aligned objects:

weights <- estimate_weights(
  b = prepared$b_estimate_table$b_estimate,
  A = prepared$ma_dosages
)

weights
sum(weights)

The output is a named numeric vector constrained to sum to 1.

Learn more

For a longer worked example, see the vignette:

vignette("how_to_use_poodleR", package = "poodleR")