| Instructor | Ellen Zhong |
| TA | Rishwanth Raghu |
| Time | Thursdays 3:00-4:20p, Computer Science Building 402 |
| TA Office hours | TBD | Slack | Link |
Recent breakthroughs in machine learning algorithms have transformed the study of proteins and other biomolecules. Deep learning algorithms designed for molecular data are advancing key scientific questions relating to molecular properties, 3D shape, interactions, and molecular design. This seminar will explore computational applications to the study of molecular systems with a focus on proteins and structural biology. We will take a holistic approach when considering problems in this domain. Students are encouraged to develop projects pursuing either classical algorithms or the latest deep learning approaches. Recommend background includes COS 324 and an introductory biology class (or a willingness to learn).
Independent work resources:
| Week | Date | Topic | Readings / Assignments |
|---|---|---|---|
| 1 | February 1 | Course overview; Introduction to machine learning in structural biology |
Additional Resources:
1. Jumper, J., Evans, R., Pritzel, A. et al. Highly accurate protein structure prediction with Alphafold. Nature 2021. 2. Tunyasuvunakool, K., Adler, J., Wu, Z. et al. Highly accurate protein structure prediction for the human proteome. Nature 2021. 3. AlphaFold2 slides. [CASP14 talk] [Michael Figurnov slides] 4. Mohammed AlQuraishi's blog post. [Link] |
| 2 | February 8 | Flash talks |
Flash talk info and sign up spreadsheet
Upload slides here |
| 3 | February 15 | Guest Speaker: Zeming Lin, Facebook AI Research |
Pre-lecture questions form
Papers: 1. Rives et al. Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences. PNAS 2021. 2. Rao et al. MSA Transformer. ICML 2021. 3. Meier et al. Language models enable zero-shot prediction of the effects of mutations on protein function. NeurIPS 2021. 4. Hsu et al. Learning inverse folding from millions of predicted structures. ICML 2022. 5. Lin et al. Evolutionary-scale prediction of atomic-level protein structure with a language model. Science 2023. |
| 4 | February 22 | Project proposal talks | |
| - | February 22 | ||
| 5 | February 29 | Project proposal talks | |
| 6 | March 7 | Project updates | |
| - | March 9 | ||
| 7 | March 14 | No class - spring recess | |
| 8 | March 21 | Project updates | |
| 9 | March 28 | Project updates | |
| 10 | April 4 | Project updates | |
| 11 | April 11 | Project updates | |
| - | April 14 | ||
| 12 | April 18 | Project presentations |
Evaluating the Efficacy of Rigid and Flexible Protein-Protein Docking Models Across Ligand Receptor Features
Your Beef is not with Shake Shack, it's with Alpha-gal: Decoding the Molecular Mechanisms Behind Alpha-gal Syndrome Predicting Protein-Protein Interactions using Graph Neural Networks Optimizing Protein Structure Analysis: A Point Cloud Approach for Cryo-EM Integrating Curriculum and Anti-Curriculum Learning into CryoDRGN Designing RNA Sequences Using Transfer Learning from ProteinMPNN Characterizing Distribution Shift in Protein Language Models with Adversarial Datasets Predicting RNA Conformational Ensembles with Deep Learning Exploring the Predictive Potential of ESM-1b on Identifying Sub-Cellular Protein Localizations Impact of Myostatin Mutations on Muscle Growth: Insights from Docking Simulations |
| 13 | April 25 | Project presentations | |
| - | April 28 | ||
May I partner with someone? Yes, you may. However, please make sure that each of the partners has a clear area of responsibility. Each person will have to create their own proposal, presentation, and paper. Some overlap in the area of prior research is understandable.