Hap-Ne snakemake

HapNe is a tool for haplotype-based inference of recent effective population size. To this day, HapNe is only available for human datasets, as it requires the input data to come from a human reference genome.

This repository contains a snakemake workflow that adapts HapNe to work with non-human datasets. Note that I am not the author of HapNe.

Reference: Haplotype-based inference of recent effective population size in modern and ancient DNA samples

TL;DR

You can run the two modes of HapNe (LD and IBD) with a toy example by running the following commands:

snakemake --use-conda -c4 --configfile example/config_ld.yaml
snakemake --use-conda -c4 --configfile example/config_ibd.yaml

Installation

To install the workflow, clone the repository and install the required packages:

mamba create -n hapne-snakemake -f conda_environment.yaml
conda activate hapne-snakemake

Alternatively, you can rely on the --use-conda to automatically create the environment if you have snakemake installed.

HapNe-LD

Input data

The input data for HapNe-LD is a series of VCF files and genetic maps in SHAPEIT format. Every pair of VCF and genetic map should have the same prefix and correspond to a chromosome (or chromosome arm). An example of the expected input data is in the example directory.

Configuration

The pipeline is configured through a config.yaml file. A minimal configuration file is shown below. Additional options can be added such as apply_filter or nb_bootstraps. Refer to the HapNe documentation for more information.

data_dir: "example/"
out_dir: "results/"
map_file_suffix: ".shapeit.map"
method: "ld"

The exampledirectory should contain the input data. The suffix of the compressed VCF files should be .vcf.gz. The suffix of the genetic map files (in shapeit format) can be customized in the config.yaml file. Intermediate files will be stored in the steps directory. The output of the pipeline is stored in the results directory and consists of (1) the estimated haploid effective population size (*_hapne_estimate.csv), (2) a summary of the inference (*_hapne_summary.csv), (3) a plot of the residuals (*_hapne_residuals.png), and (4) a plot of the estimated effective population size (*_hapne_pop_trajectory.png).

Running the pipeline

To run the pipeline, execute the following command:

snakemake -c8 --configfile config.yaml

HapNe-IBD

Input data

The input data for HapNe-LD is a series of phased VCF files and genetic maps in plink format. Every pair of VCF and genetic map should have the same prefix and correspond to a chromosome (or chromosome arm).

Additionally, the pipeline requires a TSV file that defines the regions of all analyzed chromosomes (see example/regions.tsv). This file should contain the columns 'CHR' (chromosome identifier), 'FROM_BP' (start of the region), 'TO_BP' (end of the region), 'NAME' (name of the region), and 'LENGTH' (length of the region in cM).

Configuration

The pipeline is configured through a config.yaml file. A minimal configuration file is shown below. Additional options can be added such as apply_filter or nb_bootstraps. Refer to the HapNe documentation for more information.

data_dir: "example/"
out_dir: "results/"
map_file_suffix: ".plink.map"
genome_build: "example/regions.tsv"
method: "ibd"
# IBD-detection parameters
hapibd_flags: ""
# Post-processing parameters
gap: 0.6 # Maximum gap between two IBD segments to merge them (in cM)
discord: 1 # at most one discordant homozygote

The exampledirectory should contain the input data. The suffix of the compressed VCF files should be .vcf.gz. The suffix of the genetic map files (in plink format) can be customized in the config.yaml file. Intermediate files will be stored in the steps directory. The output of the pipeline is stored in the results directory and consists of (1) the estimated haploid effective population size (*_hapne_estimate.csv), (2) a summary of the inference (*_hapne_summary.csv), (3) a plot of the residuals (*_hapne_residuals.png), and (4) a plot of the estimated effective population size (*_hapne_pop_trajectory.png).

The pipeline will use HapIBD to detect IBD segments and process them according to the suggestions of the HapNe authors. The hapibd_flags parameter can be used to pass additional flags to HapIBD. The gap and discord parameters are used to merge IBD segments and remove discordant homozygotes, respectively, using the merge-ibd-segments tool.

Running the pipeline

To run the pipeline, execute the following command:

snakemake -c8 --configfile config.yaml

Analyses of ancient samples

I haven't looked into this, but feel free to make a pull request if you give it a try! Of course, if you have ancient human samples, you should use the original HapNe tool.