Source code and data for Russian grape parentage study

This repository contains source code and data relating to "SNP-based analysis reveals authenticity and genetic similarity of Russian indigenous grape cultivars" article.

Source code

genotype-scaling.py

The script performs multidimensinal scaling with tSNE implementation from scikit-learn or UMAP (Uniform Manifold Approximation and Projection) https://umap-learn.readthedocs.io/en/latest/. It requires genotypes in tab-separated table and metadata, including 1 or 2 user-defined categories showing as colors and different characters on the plots.

genotype-analysis.py

This script implements all-in-one analysis: IBS-distance calculation, Opposing homozygotes counting, finding of possible parent-offspring relations and parentage trios, plotting dendrograms based on IBS-distance and 2D-representation of IBS-distance data using tSNE.

Requirements

• numpy
• pandas
• umap
• matplotlib
• seaborn
• sklearn
• scipy

The packages may be easilly installed with pip or conda.

Usage

-h, --help
show the help message and exit

-i INPUT, --input INPUT
Genotype in tab-separated table.

-m METADATA, --metadata METADATA
Tab-separated metadata for genotypes, if available. First column contains identifiers and should match genotype IDs from first row of input datafile.

-s SHIFT, --shift SHIFT
Number of columns between first column and columns with genotypes.

-a {umap,tsne}, --algo {umap,tsne} Algorithm for dimensionality reduction: umap (default) or tsne.

-p PERPLEXITY, --perplexity PERPLEXITY
tSNE perplexity parameter.

map-and-call.sh

Bash script for sequencing reads mapping onto reference (12X), 527 SNPs calling and genotypes extraction.

Requirements

• bowtie2
• samtools
• bcftools
• wget

Usage

./map-and-call.sh FORWARD_READS REVERSE_READS

Data and metadata

snp.data.tsv and metadata.tsv

Metadata (geographical group and berry colour) and genotypes of studied V. vinifiera cultivars coupled with data from Laucou et al., 2018. Both tables are tab-separated.

References

McInnes, L., Healy, J., & Melville, J. (2018). Umap: Uniform manifold approximation and projection for dimension reduction. arXiv preprint arXiv:1802.03426.

Van Der Maaten, L. (2014). Accelerating t-SNE using tree-based algorithms. The Journal of Machine Learning Research, 15(1), 3221-3245.

Laucou, V. et al. (2018). Extended diversity analysis of cultivated grapevine Vitis vinifera with 10K genome-wide SNPs. PloS one, 13(2), e0192540.