ETH Master’s student
Drug interaction modeling for prostate cancer
To remedy the ever-rising costs for the production of novel pharmaceutical drugs ("Eroom's Law"), better in silico modeling strategies are needed. In the last decade, advances in high-throughput screening and deep learning have enabled the development of large-scale, multimodal predictive models for virtual drug screening (Perez-Sianes et al., 2019). These neural networks are typically bimodal, based on convolutional or attention components and predict the interaction of a drug with other entities, such as proteins, cells or other drugs. Along these lines, previous work in our group (Manica et al., 2019, Cadow et al., 2020) has led to the development of PaccMann, an interpretable, deep-learning based framework that predicts anticancer drug sensitivity.
This master thesis project is a collaboration between IBM Research Europe in Zurich and the Urogenus laboratory at the University of Bern. The primary objective is to develop drug interaction models for prostate cancer, a cancer type that still lacks effective long-term and targeted treatment. The student will extend our previous models (Manica et al., 2019, Born et al, 2021, Weber et al., 2021), and apply them to prostate cancer by exploiting large-scale data from cancer cell lines (e.g., GDSC), and state of the art in vivo models (PDXs/PDOs). The ultimate goal is to identify effective, targeted and personalized treatments for selected prostate cancer patients, based on the individual multi-omic tumor profiles. These efforts will be focused on drug repurposing strategies, i.e., using predictive models to screen databases of available drugs. We will evaluate the developed complex architectures by benchmarking them against more simple predictors, such as kNNs as non-parametric alternatives on several multimodal drug interaction tasks. Based on the student's progress and interest, emphasis will be optionally given on developing and applying deep conditional generative models for de-novo molecular design (Born et al., 2021) and propose novel, targeted drugs.
We invite applications from ETH Master students only with a background in Computer/Data Science, Computational Biology/Bioinformatics or related fields. The ideal candidate should have a solid background in machine learning, deep learning and data analysis. Strong programming skills in Python and practical experience with PyTorch or at least one other deep-learning framework (Tensorflow, Keras) are essential. Prior knowledge of molecular biology is not a prerequisite.
IBM is committed to diversity at the workplace. With us you will find an open, multicultural environment. Excellent flexible working arrangements enable all genders to strike the desired balance between their professional development and their personal lives.
How to apply
If you are interested in this position, please submit your application including a recent curriculum vitae.
- Manica, Matteo, et al. "Toward explainable anticancer compound sensitivity prediction via multimodal attention-based convolutional encoders." Molecular Pharmaceutics 16.12 (2019): 4797-4806.
- Cadow, Joris, et al. "PaccMann: a web service for interpretable anticancer compound sensitivity prediction." Nucleic Acids Research 48.W1 (2020): W502-W508.
- Born, Jannis, et al. "PaccMannRL: De novo Generation of Hit-like Anticancer Molecules from Transcriptomic Data via Reinforcement Learning." iScience (2021): 102269.
- Born, Jannis, et al. "Data-driven Molecular Design for Discovery and Synthesis of Novel Ligands-A case study on SARS-CoV-2." Machine Learning: Science and Technology (2021).