https://github.com/abailoni/predicting-apmaldi-response

https://github.com/abailoni/predicting-apmaldi-response

Last synced: 24 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/abailoni/predicting-apmaldi-response
• Owner: abailoni
• License: mit
• Created: 2022-06-14T07:49:44.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2023-04-06T12:50:28.000Z (almost 2 years ago)
• Last Synced: 2024-11-05T11:55:25.398Z (2 months ago)
• Language: Jupyter Notebook
• Size: 175 MB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Predicting AP-MALDI response using Machine Learning

The trained models and predictions can be found in the [training_results/paper_results](./training_results/paper_results) folder.

### How to Install

- `conda create --name apMALDIresponseEnv python=3.8`

- `conda activate apMALDIresponseEnv`

- `pip install -r requirements.txt`

- `python setup.py install`

### Predicting AP-MALDI response for custom molecules

Follow the instructions in the notebook [predict_intensities.ipynb](notebooks/predict_intensities.ipynb) to predict AP-MALDI response on your custom set of molecules.

### Reproducing paper results

To reproduce the paper results, follow these steps:

- Retrain the models by running `python train_models.py --exp_name `

- After training is done, models and predictions can be found in the `training_results/YOUR_NEW_EXPERIMENT_NAME` folder

- To compute scores and make plots, use the [evaluate_trained_models.ipynb](notebooks/evaluate_trained_models.ipynb) notebook