Data Science Meets Biology

Hello everyone!

We're excited to announce that registration for the "Data Science Meets Biology" Hackathon is now open! This groundbreaking event is organized by MDC PhD students, in collaboration with Bayer, a global leader in pharmaceuticals and biotechnology.

Format: Hybrid format – in-person at Bayer's office on September 6 and 9, with virtual sessions over the weekend.

About the Hackathon:

This is your chance to tackle real-world healthcare challenges by merging the power of Biology and Data Science. Whether you're a biologist, a data scientist, or someone passionate about healthcare innovation, your expertise and creativity are invaluable here.

What to Expect:

Collaborative Problem-Solving: Work with brilliant minds on impactful projects.
Expert Mentorship: Get guidance from industry leaders.
Networking: Connect with professionals and expand your network.
Exciting Prizes: Win awards for the most innovative solutions!

Mentors:

• Matthias Orlowski & Marc Osterland: Machine Learning Engineers at Bayer Pharmaceuticals
• Jonas Münch: Head of Pharmacology & Safety IT
• Ni Fang: Management Support to Chief of Staff of PH R&D Data Science & Artificial Intelligence
• Udo Schubert: Data Specialist Pharma Production

Projects:

1. Build a Generative AI-Powered Scientific Knowledge Engine (By Jonas Münch)

In Bayer Pharma Research & Development, accessing insights from vast scientific data can be time-consuming. This project challenges you to create a chatbot using modern Large Language Models (e.g., GPT-4) and at least 5 research papers of your choice. Develop a fully functional Retrieval Augmented Generation (RAG) web application using AWS, enabling users to query the papers and receive accurate answers with references.

2-3. Apply Data Science/AI for Performance Analysis of an innovative Manufacturing Process (two projects) (By Udo Schubert)

In a new Product Supply facility, achieving uninterrupted normal production is challenging due to the need to learn new procedures without extensive test runs. This project involves analyzing datasets from real production runs at a Bayer facility using Data Science/AI methods. Your goal is to derive insights to enhance production efficiency and reduce batch duration by addressing key factors causing interruptions.

4. Boosting phenotypical profiles for early drug discovery (By Marc Osterland  & Matthias Orlowski)

Cell painting and morphological profiling are key in early drug discovery for evaluating the effects of compounds on cellular phenotypes. This project focuses on identifying and removing batch effects from morphological feature vectors derived from cell painting microscopy images. Your goal is to preprocess the data, eliminate batch effects, and enhance the detection of meaningful phenotypical clusters.

5. Sensitivity and uncertainty of machine learning model (By Ni Fang)

Discovering and developing pharmaceutical products requires a profound understanding of therapeutic agents and their properties at the molecular level. This project aims to estimate the thermodynamic stability of proteins, such as antibodies or enzymes, using machine learning. You'll assess protein stability with respect to mutations and use uncertainty estimates to suggest mutations for improved stability.

Registration deadline: 31.08.2024

Contact Details:

Bogdan Avanesy: bogdan.avanesyan@mdc-berlin.de
Ekaterina Vasileva: ekaterina.vasileva@mdc-berlin.de