Mini-project
Pharma & Life Sciences
Protein function classifier from sequences
Predict a protein's family or function from its amino-acid sequence.
The brief
A research lab sequences proteins faster than it can study them. Build a classifier that predicts a protein's family or function from its sequence, starting with simple sequence features and moving to a pre-trained protein language model. Good looks like: a clear gain from the language model over simple features, and an analysis of which families are hard.
What you will build
- Download sequences and labels for a few protein families
- Build k-mer features and a baseline classifier
- Use embeddings from a pre-trained protein model
- Compare results per family
- Explain the biology behind the hard cases
Free, open data to build it with
- UniProt protein database ↗ UniProt Consortium
- RCSB Protein Data Bank ↗ RCSB PDB
Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.