Qst–Fst Comparative Analysis under Neutral and Quantitative Trait Simulations

Overview

This repository contains the full pipeline used to simulate, process, and analyze genetic data for comparing Qst and Fst estimates under different population structures and genetic architectures. The workflow combines:

• Forward-time simulations (SLiM)
• Tree sequence processing (Python)
• Quantitative trait analysis (R)

The repository provides:

• SLiM scripts for simulating island and stepping-stone models.
• Python scripts to process simulated tree sequences into VCFs with added mutations and founder sampling.
• R code implementing the Guillaume–Whitlock (GW) framework for Qst estimation and statistical testing.
• Shell scripts for automated and parallelized execution on HPC clusters (SLURM).

---

Repository Structure

├── scripts/ # Core pipeline code

│ ├── Full_code_analysis_Qst_Fst.Rmd # Analysis and visualization of final results

│ ├── GW_Qst_Fst.r # Main algorithm (Guillaume–Whitlock framework) adapted by Jikaël Ntoko

│ ├── process_stepping_tree.py # Tree → VCF (stepping-stone model)

│ ├── process_tree_island.py # Tree → VCF (island model)

│ ├── run_batch_vcf.sh # Parallel VCF processing launcher

│ ├── run_one_vcf.sh # Single execution of tree → VCF

│ ├── run_one_seed.sh # Single SLiM simulation run

│ ├── run_slim_batch.sh # Array job: multiple SLiM simulations

│ ├── run_one_qst.sh # Single Qst–Fst analysis (R)

│ ├── run_qst_fst.sh # Batch Qst–Fst analysis (SLURM array)

│

├── slim_script/ # Forward simulations in SLiM

│ ├── island_neutral.slim # Island model simulation

│ ├── stepping_neutral.slim # Stepping-stone model simulation

I

Installation

Requirements

• SLiM 4+ (forward-time population genetic simulator)
• Python 3.9+ with:
  • tskit, pyslim, msprime, numpy, pandas
• R (≥ 4.2) with:
  • tidyverse, hierfstat, JGTeach, vcfR, gaston, VGAM, boot

This pipeline is designed for HPC clusters running SLURM; adaptation may be needed for local execution

# Python environment
conda create -n qstfst python=3.10 numpy pandas pyslim tskit msprime
conda activate qstfst

# R environment (example)
module load R/4.3.1
Rscript -e 'install.packages(c("tidyverse","vcfR","boot","VGAM"))'

Usage

1. Run Simulations

Island model example:

slim slim\_script/island\_neutral.slim

Stepping-stone model:

slim slim\_script/stepping\_neutral.slim

Or submit arrays on SLURM:

bash scripts/run\_slim\_batch.sh

2. Process Tree Sequences → VCF

bash scripts/run\_batch\_vcf.sh

3. Qst–Fst Analysis

Run a single replicate:

bash scripts/run\_one\_qst.sh

Or submit in parallel:

bash scripts/run\_qst\_fst.sh

4. Visualization and Results

Open the R Markdown file for figure generation and summary statistics:

rmarkdown::render("scripts/Full\_code\_analysis\_Qst\_Fst.Rmd")

Authors & Affiliations

• Author: Jikaël Ntoko, University of Lausanne (UNIL)
• Contributor (algorithmic component): Isabela Do’Ò (original algorithm (Whitlock and Guillaume implementation))