I'm trying to delve deep into the world of cancer genomics and explore the potential of AI in medicine. This 1st project of hopefully many is all about predicting how cancer cells react to different drugs based on their genetic makeup. I will use these projects to explore computational drug discovery while pursuing my MSc in Precision Cancer Medicine at Oxford and leveraging the skills I have learnt in my BSc in Computer Science.
Hope you enjoy, if you have any feedback or questions then please reach out to me here: LinkedIn
- Learning & Exploration: I wanted to get hands-on with medical datasets and understand the intricacies of combining them with AI.
- Personal Growth: Not only does this help build a robust portfolio and gain practical experience, but I wanted to gain the skills neccessary to become a Machine Learning Engineer in Cancer Medicine.
- Social Good: The potential to influence personalized cancer treatments is impactful and should I one day enter this career pathway I can only hope it will benefit the masses with cancer therapies.
- Data Analysis: Comprehensive analysis of GDSC along with data exploration and cleaning if need be.
- Modeling: Using Python, Scikit-learn, and TensorFlow to build predictive models.
- Visualization: Interactive dashboards built with Gradio and Streamlit for real-time predictions and accessibility.
- Testing: Unit tests to ensure model robustness, accuracy and data consistency.
- Prerequisites: Ensure you have Python, Scikit-learn, TensorFlow, Gradio, and Streamlit installed.
- Clone the Repository:
git clone [your-repo-link]
- Navigate to Directory:
cd [your-repo-name]
- Install Dependencies:
pip install -r requirements.txt
- Run the App:
streamlit run app.py
- Genomics of Drug Sensitivity in Cancer: Sourced from GDSC
Feel free to fork the project, open a PR, or suggest any enhancements. If you found any bugs, please do let me know!
- Inspired by projects like DeepChem, Chanin Nantasenamat and others in the AI-medical field.
- Special thanks to the Wellcome Sanger Institute for providing the datasets.
- Grateful for the open-source community for the plethora of resources available.
This project is licensed under the MIT License - see the LICENSE file for details.