TimeAdapt

TimeAdapt makes (will eventually make) a joint inference of demography and selection from longitudinal population genomic data. The inference is based on approximate Bayesian computation via random forest. It takes whole genome data and information about the ages of samples to perform simulations (using SLiM + pyslim/msprime).

Requirements

TimeAdapt is a collection of scripts in Python and SLiM. They have been tested in an Ubuntu (20.04) machine using a Conda environment and using a Snakemake workflow to run them. The Conda environment was created with the following commands (see file timeadapt.yml to get version number of each package):

conda create -n timeadapt python==3.8.10 r-base=3.6.3
conda activate timeadapt
pip install tskit==0.4.1
pip install msprime==1.1.0
pip install pyslim==0.700
pip install scikit-allel==1.3.5
pip install pandas
pip install scipy
pip install pytest
pip install flake8
conda install slim=3.7
conda install -c r r-rcarbon=1.2.0
conda install -c r r-ini
conda install -c r r-extraDistr
conda install -c r r-testthat
conda env export > timeadapt.yml

Reminder for creating environment from file:

conda env create -f timeadapt.yml

Install abcrf in R via install.packages() with appropriate path, e.g.:

install.packages('abcrf','/home/USER/anaconda3/envs/timeadapt/lib/R/library').

Usage

Input files:

• config file: a ini file with options regarding the Setting, Model, Prior and Statistics to be used in your analysis

• sample file: a text file (in the form of a space separated table) with metadata of your samples (ID, age, sequencing coverage, etc)

• genome file: a text file with a description of the genome organization (chromosomes, recombination map)

• data file: a vcf file with the genetic data

To run different parts of the analysis with snakemake :

snakemake RULE -C options_file='PATH/TO/YOUR/CONFIG_FILE.ini'

Where RULE is one of the rules defines in the snakefile. For instance, running snakemake reftable -C options_file='tests/config_project.ini' will create small reference table using parameters in file tests/config_project.ini. Typically the user will use rule reftable to run simulations and create the reference table (parameters, summary statistics and latent variables) and abc to grow random forests and perform approximate Bayesian computation inferences (rule abc will call reftable if there is no reference table; there is no need to do it in two steps but it can be usful in many cases).

Before running your analysis is highly recommended to performs some tests. Unit tests (using pytest) can be run using snakemake test.

Cleaning old files

You can remove all files (from a specific project or from all projects) from your results folder:

snakemake clean_project -C options_file='path/to/your/config_file.ini'
snakemake clean

Input parameters (in *.ini config file)

Parameter name	type	description
[Settings]
verbose	integer	Level of information to output on screen (0 = minimal; 1,2,3,... = increasing detailed output; -1 = none).
project_name	string	name of the analysis project, a folder with that name is created in results folder and all output written inside.
batch	integer	identifier of a batch of simulations for the project. A folder named with that identifier is created within the project folder and results from the simulations of that bacth written inside.
sample_file	string	path + file name containing sample information (see below)
genome_file	string	path + file name containing genome information (see below)
num_of_sims	positive integer	number of simulation (in the batch)
seed	integer	seed for the random number generator
[Model]
generations_forward	positive integer	number of generations simulated in forward (by SLiM)
periods_forward	positive integer	number of periods in which the forward simulation is divided
[Priors]
gen_len_prior_sh1		prior for generation length (shape1 of rescaled beta distribution)
gen_len_prior_sh2		prior for generation length (shape2 of rescaled beta distribution)
gen_len_prior_min		prior for generation length (minimum of rescaled beta distribution)
gen_len_prior_max		prior for generation length (maximum of rescaled beta distribution)
pop_size_prior_min	positive integer	prior for population size (minimum of uniform distribution)
pop_size_prior_max	positive integer	prior for population size (maximum of uniform distribution)
mut_rate_prior_mean		prior for mutation rate (mean of normal distribution)
mut_rate_prior_sd		prior for mutation rate (sd of normal distribution)

Example:

Input files

• Sample information file: Text file with information on the samples to be analysed. First line should be a header containing the following column names:

  • "sampleID": String identifying each individual. Do not use spaces or tabs.
  • "age14C": Point estimate of age from radiocarbon dating (for ancient samples, set NA for modern samples or samples with a known calendar year). It indicates time of death.
  • "age14Cerror": Error for age from radiocarbon dating.
  • "year": Calendar year for the time of sampling (for present and modern samples, such as museum/collection samples, set NA for ancient samples)
  • "coverage": Numeric value indicating sequencing depth (do NOT add "x", write "30" not "30x")
  • "damageRepair": Indicate whather the library for ancient samples used a damage repair step. Indicate TRUE or FALSE (indicate TRUE for modern samples)
  • "groups": Numerical code to indicate the groups of individuals for which summary statistics will be calculated. For each digit, individuals with the same number form one group, each digit defines a set of groups. Summary statistics comparing groups (such as Fst). Numbers indicating groups must be consecutive and start with 0.

• Genome information file:

...

How to cite TimeAdapt

TimeAdapt implements the method described by Pavinato et al. (2020) with some modifications. The code puts together several tools that should be acknowledged when using TimeAdapt: SLiM, pyslim, msprime (TODO: add full citation for these)

Funding

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 791695 (TimeAdapt).

Licence

TimeAdapt: joint inference of demography and selection

Copyright (C) 2021 Miguel de Navascués, Uppsala universitet, INRAE

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see https://www.gnu.org/licenses/.