Outlier Removal in Cryo-EM via Radial Profiles

This repository contains an implementation of outlier removal in cryo-electron microscopy (cryo-EM) datasets using radial profiles. The method focuses on removing outliers to enhance the quality of the 2D class averages and downstream analysis.

Setup

1. Create a Virtual environment:

   To set up the required dependencies, create a Virtual environment with python=3.8 using the provided requirements.txt file:

   python3.8 -m venv .venv
   source .venv/bin/activate
   pip install -r requirements.txt
   

2. Synthetic Data Experiments:

   To reproduce the synthetic data experiments, run the following script:

   python run_synthetic_experiment.py --use_pure_noise --use_part_particles --use_multi_particles
   

3. Real Data Experiments:

   To reproduce the experiments on real cryo-EM datasets, execute:

      python run_real_data_experiment.py --dataset_name "10028" --extraction_size 360 --outlier_removal
   

Requirements:

To run the real data experiments, ensure you have a directory with the following structure:

Dataset Directory:

• The directory name should match the dataset name.
  • Contains .mrc files (micrographs).
  • Contains particle picks generated by Relion.
  • Contains CTF (Contrast Transfer Function) corrections for each micrograph.

Example: An example of the required file structure and data is provided in the 10028 dataset directory.