Breed/Variety Composition Ratio Estimator

This module evaluates the breed or variety composition ratio of samples based on the SNP (Single Nucleotide Polymorphism) allele frequency information.

Two inputs must be provided to evaluate the composition ratio:

• The SNP allele frequency of the samples.
  • Must be a matrix (in R environment) with specified rownames and colnames, or in a VCF file format that stores the allele frequency of ALL samples (See Section 2 for details).
• The SNP allele frequency of the reference breed or variety (Curated from the published database, such as the BGVD database for cow breed allele frequency data)
  • Must be a matrix (in R environment) with specified rownames and colnames (See Section 2 for details).

1. Directory

There is a simple sub-repository. All input files are stored in the data directory.

• Input:
  • cowbreed_samplemix_allele_frequency.rdata: The SNP allele frequency of 16 synthetic cow breed mixture samples.
    • Stored as a variable samplemix_allele_frequency
    • Hereford, Jersey, and Holstein cow breeds are mixed in different ratios.
    • The sample label (colnames) recorded the mix ratios, for example:
      • The Label H10_J20_Hol5 means the ratio of Hereford, Jersey, and Holstein is (10 : 20 :5)
      • The Label H10_JX_Hol5 means the ratio of Hereford, Jersey, and Holstein is (10 : 0 : 5)
  • cowbreed_ref_allele_frequency.rdata: The SNP allele frequency of the reference cow breeds (Hereford, Jersey, and Holstein in this case)
    • Stored as a variable ref_allele_frequency

2. Getting Started

The code is tested under the R environment version 4.2.0 with the following packages: nnls, foreach, and doParallel. You may install them using the following command in R:

install.packages(c("nnls", "foreach", "doParallel"))

The source code file nnls_functions.R contains several functions. To use that, just source the file:

source("src/nnls_functions.R")

(Optional) Preparing the Reference Allele Frequency Data

The reference allele frequency data is a matrix with specified rownames and colnames.

• rownames: The SNP name or ID. (Note: The SNP name or ID should be consistent with the SNP name or ID in the sample allele frequency data)
• rownames: The breed (variety) names.

The allele frequency values are between 0 to 1. Here is an example (only showing the first six SNPs):

> head(ref_allele_frequency)
            Hereford_val Holstein_val Jersey_val
1:324_A/G          0.167        0.300      0.273
1:340_G/A          0.167        0.300      0.250
1:380_G/T          0.167        0.300      0.208
1:2377_A/G         0.048        0.311      0.375
1:11412_C/G        0.476        0.433      0.083
1:14259_C/T        0.500        0.500      0.458

(Optional) Preparing the Allele Frequency Data of the Mixture Samples

The allele frequency data of the mixture samples is a matrix with specified rownames and colnames.

• rownames: The SNP name or ID. (Note: The SNP name or ID should be consistent with the SNP name or ID in the reference allele frequency data)
• rownames: The sample names.

The allele frequency values are between 0 to 1. Here is an example (only showing the first six SNPs and the first five samples):

> head(samplemix_allele_frequency)
            H10_J20_Hol5 H10_J5_Hol20 H10_J5_HolX H10_JX_Hol5 H20_J10_Hol5
1:324_A/G      0.7941176    0.2000000   0.6363636   0.0000000    0.4210526
1:340_G/A      0.9090909    0.9166667   0.7272727   0.5000000    0.4324324
1:380_G/T      0.7857143    0.8604651   0.6000000   0.3333333    0.3888889
1:2377_A/G     1.0000000    1.0000000   1.0000000   1.0000000    1.0000000
1:11412_C/G    0.1481481    0.1666667   0.5714286   0.4000000    0.4000000
1:14259_C/T    0.3157895    0.2500000   0.6666667   1.0000000    0.6285714

Tools for preparing the Allele Frequency Data Matrix from the VCF file

The extract_ad_matrix_info_batch (in nnls_functions.R) can help you to get the Allele Frequency Data Matrix from the VCF file generated from the tools for SNP detection (e.g., GATK)

(Note: The VCF files must contain the allele frequencies of ALL samples, so they should be yielded from the joint-genotyping approach or combined from multiple VCF files that contain the allele frequencies of a single sample):

extract_result = extract_ad_matrix_info_batch("Batch1-gatk-haplotype-annotated.vcf.gz")
samplemix_allele_frequency = extract_result$ad_matrix

2. Composition Ratio Evaluation:

Run the function run_nnls_merge_batch() to evaluate the breed/variety composition ratio of the samples:

#Load the input data
>load('cowbreed_ref_allele_frequency.rdata')
>load('cowbreed_samplemix_allele_frequency.rdata')
>result = run_nnls_merge_batch(samplemix_allele_frequency, ref_allele_frequency)
>head(result)
             Hereford_val Holstein_val Jersey_val
H10_J20_Hol5    0.1992635   0.22503754  0.5756990
H10_J5_Hol20    0.1817231   0.69454493  0.1237320
H10_J5_HolX     0.6538215   0.04202237  0.3041561
H10_JX_Hol5     0.6180149   0.38198506  0.0000000
H20_J10_Hol5    0.5362635   0.20531296  0.2584236
H20_J5_Hol10    0.5307980   0.36143639  0.1077656

3. Authors

• ChengEn Tan @https://github.com/bigghost2054

4. Contact

For any questions, please contact us at tagkopouloslab@ucdavis.edu.

5. Citation

• (Will be ready once the work is published).

6. License

The entire project is licensed under the Apache 2.0 License. Please see the LICENSE file for details.

Written with StackEdit.