GENBAIT Reproducibility

This repository contains the results of all analyses in the GENBAIT project for bait selection in BioID experiments.

A preprint describing the method and introducing a novel benchmarking platform is available:
Kasmaeifar et al. (2024) Computational design and evaluation of optimal bait sets for scalable proximity proteomics

This project is designed to be reproducible using Snakemake. Below are the instructions on how to reproduce the results of each step in the workflow using the provided configuration files.

---

Requirements

Before running the workflow, ensure you have the following installed:

• Python 3.10+
• Snakemake
• Git LFS

---

Git Installation

To install GENBAIT, you first need Git. Follow the instructions below to install Git on your system.

For Windows

1. Download the Git installer:

   • Go to the official Git website: https://git-scm.com/download/win.
   • Download the latest installer for Windows.

2. Run the installer:

   • Locate the downloaded file and double-click to open the installer.
   • Follow the prompts in the setup wizard. You can keep the default options or customize the installation.

For macOS

1. Install Git using Homebrew:

   • If you have Homebrew installed, open the Terminal and run:

     brew install git

   • Homebrew will handle the download and installation of Git.

2. Verify the installation:

   • In Terminal, type the following command and press Enter:

     git --version

   • You should see a Git version number, confirming that Git is installed.

---

Build Tools Installation

Some packages used in this repository (e.g., shap, xgboost, leidenalg) require compilation and system build tools.

For Windows

1. Download Microsoft C++ Build Tools:
   https://visualstudio.microsoft.com/visual-cpp-build-tools/

2. In the installer, select the following:

   • C++ build tools workload
   • MSVC v14 or later
   • Windows 10 or 11 SDK

3. Complete the installation and restart your terminal.

If these tools are missing, you may encounter errors such as:
error: Microsoft Visual C++ 14.0 or greater is required

---

For macOS

1. Install Xcode Command Line Tools:
   Open Terminal and run:

   xcode-select --install
   

Setup

Create a Virtual Environment

It is recommended to create a virtual environment to manage dependencies:

python -m venv genbait_env
source genbait_env/bin/activate  # On Windows use `genbait_env\Scripts\activate`

Install the Package

Navigate to the root directory of the project and run:

git clone https://github.com/camlab-bioml/genbait_reproducibility.git
cd genbait_reproducibility
pip install .

This will install the package along with all required dependencies.

Running the Workflow

To reproduce the results for each dataset, run the Snakemake workflow. The configuration files for each dataset are located in the config/ directory.

# Example: Load data step for dataset1 using all available CPU cores
snakemake --cores all load_data --config dataset=dataset1

# 1. Load data
snakemake --cores 1 load_data


# 2. GENBAIT evaluation

# Run the Genetic Algorithm (GA)
snakemake --cores 1 run_ga

# Evaluation
snakemake --cores 1 ga_evaluation

# Run GENBAIT for different bait lenghts and seeds
snakemake --cores 1 ga_number_of_baits_seeds

# Bait lengths and seeds evaluation
snakemake --cores 1 seeds_evaluation


# 3. Machine learning feature selection

# Run ML Methods
snakemake --cores 1 run_ml_methods

# Plot ML Methods
snakemake --cores 1 plot_ml_methods


# 4. NMF metrics rules

# Mean NMF correlation
snakemake --cores 1 plot_nmf_scores

# Min NMF correlation
snakemake --cores 1 plot_nmf_scores_min

# Mean NMF Cosine similarity
snakemake --cores 1 plot_nmf_cos_scores

# Min NMF Cosine similarity
snakemake --cores 1 plot_nmf_cos_scores_min

# Mean NMF KL divergence
snakemake --cores 1 plot_nmf_kl_scores

# Max NMF KL divergence
snakemake --cores 1 plot_nmf_kl_scores_min

# NMF ARI
snakemake --cores 1 plot_nmf_ari_scores

# Min NMF purity score
snakemake --cores 1 plot_nmf_ari_scores_min

# Mean NMF Jaccard GO index
snakemake --cores 1 plot_nmf_go_scores

# Min NMF Jaccard GO index
snakemake --cores 1 plot_nmf_go_scores_min


# 5. Non-NMF metrics rules

# Remaining preys percentage
snakemake --cores 1 remaining_preys_evaluation

# GO retrieval percentage
snakemake --cores 1 go_evaluation

# Leiden ARI
snakemake --cores 1 leiden_evaluation

# GMM ARI
snakemake --cores 1 gmm_hard_evaluation

# Mean GMM correlation
snakemake --cores 1 gmm_evaluation

# Combined metrics plot
snakemake --cores 1 combined_metrics


# 6. Other analyses rules 

# Topology analysis
snakemake --cores 1 topology_analysis

# Runtime analysis
snakemake --cores 1 runtime_analysis

# Individual components correlation
snakemake --cores 1 individual_components_correlation


# 7. Dataset1-specific analysis

# Bait expression analysis
snakemake --cores 1 bait_expression_analysis

# Simulation expression analysis
snakemake --cores 1 simulation_expression_analysis

# 8. Cobined datasets plots

# Comined mean NMF correlation
snakemake --cores 1 combined_nmf_corr

# Comined min NMF correlation
snakemake --cores 1 combined_nmf_min_corr

# Comined mean NMF Cosine similarity
snakemake --cores 1 combined_nmf_cos

# Comined min NMF Cosine similarity
snakemake --cores 1 combined_nmf_min_cos

# Comined mean NMF KL divergence
snakemake --cores 1 combined_nmf_kl

# Comined min NMF KL divergence
snakemake --cores 1 combined_nmf_min_kl

# Comined NMF ARI
snakemake --cores 1 combined_nmf_ari

# Comined min NMF purity score
snakemake --cores 1 combined_nmf_min_purity

# Comined mean NMF go
snakemake --cores 1 combined_nmf_go

# Comined min NMF go
snakemake --cores 1 combined_nmf_min_go

# Combined remaining preys
snakemake --cores 1 combined_remaining_preys

# Combined GO retrieval 
snakemake --cores 1 combined_go_retrieval

# Combined Leiden ARI
snakemake --cores 1 combined_leiden

# Combined GMM ARI
snakemake --cores 1 combined_gmm_hard

# Combined mean GMM correlation
snakemake --cores 1 combined_gmm_correlation

# 9. Final step: mark workflow as completed
snakemake --cores 1 finalize_workflow