AlleleDoser

Germline variant analysis with copy number and phased genotype integration

Description

This project provides an R script for analyzing germline variants in tumor samples by integrating germline variant counts, copy number data, metadata (e.g., tumor purity), and phased genotypes. The script calculates expected allele frequencies, confidence intervals, and final allele dosages, incorporating both copy number variation (CNV) and phased genotype information.

Installation

Clone the repository and ensure you have the necessary R packages:

# Clone the repository
git clone https://github.com/HautaniemiLab/AlleleDoser.git
cd AlleleDoser

# Install required R packages (if not already installed)
Rscript -e "install.packages(c('tidyverse', 'Rsamtools'))"

Usage

Run the script as follows:

Rscript alleleDoser.R \
  path/to/germline_variants.tsv \
  path/to/segment_file.tsv \
  path/to/metadata.tsv \
  path/to/output_directory \
  patientID \
  path/to/normal_samples.txt \
  path/to/phased_genotypes.bcf

Inputs

The script requires seven command-line arguments:

1. Germline in tumor genotype
2. Segment file
3. Metadata file
4. Output directory
5. Patient ID
6. Normal sample IDs
7. Phased genotypes

Input File Descriptions

1. Germline in Tumor Genotype File (TSV)

You can generate the input TSV file from your VCF file using the following GATK command:

gatk VariantsToTable \
  -V input.vcf \
  -O output.tsv \
  -F CHROM -F POS -F REF -F ALT -GF AD

Note: The VCF file should contain a normal sample along with the tumor samples.

Example Output Table Format:

CHROM	POS	REF	ALT	patient1_sample1	patient1_sample2	patient1_sample3
chr1	12345678	A	G	10,5	8,7	12,3
chr2	87654321	C	T	20,2	15,4	18,6

---

2. Segment File (TSV)

The segment file contains CNV information for each sample.

Required Columns:

• sample: Sample ID.
• chr: Chromosome identifier.
• startpos: Start position of the CNV segment (genomic coordinate).
• endpos: End position of the CNV segment (genomic coordinate).
• nMinor: Minor allele copy number within the segment.
• nMajor: Major allele copy number within the segment.

Example Segment File Format:

sample	chr	startpos	endpos	nMinor	nMajor
patient1_sample1	chr1	1000000	2000000	1	2
patient1_sample1	chr1	2000001	3000000	0	3
patient1_sample2	chr2	500000	1500000	1	1
patient1_sample3	chr2	1500001	2500000	2	2

---

3. Metadata File (TSV)

The metadata file provides sample-specific tumor purity values.

Required Columns:

• sample: Sample ID.
• purity: Tumor purity value (a numeric value between 0 and 1, representing the proportion of tumor cells in the sample).

Example Metadata File Format:

sample	purity
patient1_sample1	0.75
patient1_sample2	0.60
patient1_sample3	0.85

---

4. Output Directory

The output directory specifies the location where the final results (e.g., CSV files containing allele dosages and frequency calculations) will be saved.

• The directory must already exist before running the script.
• The path should be provided as an absolute or relative path.
• Ensure the directory has write permissions for saving output files.

---

5. Patient ID

The Patient ID is used to filter sample-specific data from the input files and name the output file (<patientID>.csv).

• The provided ID must be included in the sample names associated with that patient.
• Sample names are assumed to follow a structure like patient1_sample1, patient1_sample2, patient1_sample3, etc., where patient1 is the patient ID.

---

6. Normal Samples File (TXT)

The normal samples file provides a list of normal (non-tumor) sample IDs.

• Plain text file (.txt), with one normal sample ID per line.
• Sample IDs should match those used in the germline variant and segment files.

---

7. Phased Genotype File (BCF)

The phased genotype file contains phased genotype information for each variant.

• Binary Call Format (.bcf) file, which should be indexed.

Output

The script generates a CSV file named <patientID>.csv in the specified output directory.

Output Column Descriptions

Column	Description
sample	Sample ID.
CHROM	Chromosome identifier.
POS	Genomic position of the variant.
REF	Reference allele.
ALT	Alternate allele.
Mutation	Combined variant notation (CHROM:POSREF>ALT).
AD.0	Reference allele read count.
AD.1	Alternate allele read count.
DP	Total read depth (AD.0 + AD.1).
AF	Allele frequency of the alternate allele (AD.1 / DP).
purity	Tumor purity from the metadata file.
nMajor	Major allele copy number.
nMinor	Minor allele copy number.
totalCN	Total copy number (nMajor + nMinor).
expMajorAF	Expected alternate allele frequency if the major allele is mutated.
expMajorCI.lo	Lower bound of the 95% confidence interval for expMajorAF.
expMajorCI.hi	Upper bound of the 95% confidence interval for expMajorAF.
expMajor.pbinom.extreme	Probability of observing the alternate allele count or more extreme under the major allele hypothesis.
expMinorAF	Expected alternate allele frequency if the minor allele is mutated.
expMinorCI.lo	Lower bound of the 95% confidence interval for expMinorAF.
expMinorCI.hi	Upper bound of the 95% confidence interval for expMinorAF.
expMinor.pbinom.extreme	Probability of observing the alternate allele count or more extreme under the minor allele hypothesis.
ALTallele	Preliminary classification of which allele (major or minor) carries the mutation.
normal.AD.0	Reference allele read count in normal samples.
GT	Phased genotype.
germline	Germline status classification (0homo, 1homo, hetero).
majorfirst	Boolean indicating if the major allele is likely inherited first based on phased data.
excl_mean	Excluded rolling mean of majorfirst_numeric within a window.
phasesegmented	Final phased segment classification (TRUE/FALSE) indicating phase-consistent segments.
ALTallelefinal	Final classification of the mutated allele after incorporating phased genotype information.
ALTdosagefinal	Final dosage estimate of the alternate allele accounting for copy number and phasing.

Contributors

• Yilin Li
• Samuel Leppiniemi

License

This project is licensed under the MIT License. See the LICENSE file for details.