This repository contains the code for reproducing simulations from Williams et al. (2023), Testing Times: Challenges in Disentangling Admixture Histories in Recent and Complex Demographies.
Below are details regarding the two simulation project sub-directories (simple_models, and eurasia_model) that each contain Snakemake pipelines to execute them.
- simple_models/: contains the Snakemake pipeline to simulate the simple demographic model simulations (Models 1-4) that explore the topological and admixture parameter space.
- eurasian_model/: contains the Snakemake pipeline to simulate from the Eurasian complex demographic model.
The Snakemake pipelines were executed from a base conda environment using anaconda Command line client (version 1.7.2), Snakemake v.3.13.3, and Python 3.6.13 :: Anaconda, Inc. When running the pipeline for the first time, the Snakemake software will initialize a conda environment for each sub-project. The file, msprime-env.yaml, located in each sub-directory, contains all the software necessary to execute the simulations and Snakemake will read from this file to set-up the conda env (note conda can be finicky when transferring across operating systems, the pipeline was setup on a Red Hat Enterprise Linux 7 env.). The only other software required is the ADMIXTOOLS2 R package (https://uqrmaie1.github.io/admixtools/articles/admixtools.html) (note this has to be located in the R library available to your Snakemake/conda session).
- src/: contains the python, and r code.
- config/: contains the simulation configuration files for each demographic model. See the config/config_msp_dtwf_model[1:4].yml for details regarding the input simulation parameters.
The simple_models root directory contains the executable Snakemake pipeline smk script snakemake_msp_dtwf__qpAdm_f3_SimpleDemog.smk which you execute through:
snakemake -s snakemake_msp_dtwf__qpAdm_f3_SimpleDemog.smk --configfile config/config_msp_dtwf_model[*insert*].yml --use-conda
where you define the demographic Model you want to run, config_msp_dtwf_model[1:4].
You can override the configuration file through the command line argument --config yourparam=X. As such, to perform a single replicate small chromosome test on Model 1 run the following from the terminal:
snakemake -s snakemake_msp_dtwf__qpAdm_f3_SimpleDemog.smk --configfile config/config_msp_dtwf_model1.yml --use-conda --config simulation_model_name="Model1_Test" chromosome_length=1e6 recombination_rate=1e-8 window_length_Fstats=1e5 N_replicates=1
- src/: contains the bash, python, and r code.
- config/: contains the simulation parameters configuration code. See the config/config_msp_dtwf_eurasia.yml for details regarding the input simulation parameters.
- input/: contains the demes demography, sample sheet, and where the eig aDNA datasets need to be stored (see below).
- dev_RLibs/: contains data.table v.1.14.9 dev version to access the fwrite function, and gtools Version: 3.9.4.
The eurasia_model root directory contains the executable Snakemake pipeline smk script snakemake_msp_dtwf_ts_mu_eigaDNA__eurasia_qpAdm_f3.smk.
To perform a test run on chromosome 22, for only a single target individual and one missingness dataset, run the following:
snakemake -s snakemake_msp_dtwf_ts_mu_eigaDNA__eurasia_qpAdm_f3.smk --configfile config/config_msp_dtwf_eurasia.yml --use-conda --config genome_length="chr22" N_replicates=1 sample_sheet=input/testrun_sampleSheet_eurasia.txt empir_aDNA_prefix=v52.2_1240K_public_less100kSNPs simulation_model_name="Test_chr22" --latency-wait 7200
To perform the full eurasia simulation run:
snakemake -s snakemake_msp_dtwf_ts_mu_eigaDNA__eurasia_qpAdm_f3.smk --configfile config/config_msp_dtwf_eurasia.yml --use-conda --latency-wait 7200
In order to perform the data missingness aDNA step please download the three Eigenstrat aDNA datasets (v52.2_1240K_public_SimsSubset_SWEurasia, v52.2_1240K_public_100k500kSNPs, v52.2_1240K_public_less100kSNPs) to the input directory which can be accessed at the following link: eig_aDNA_data_dir