Ahmed Doghri Logo Image
Ahmed Doghri

proteinmask

I wanted a bio-AI demo that stays honest. proteinmask uses toy protein-like strings, recovers conserved motifs at 0.915, and makes no wet-lab claim.

proteinmask, a toy masked protein sequence infilling benchmark

A Bio Demo With Boundaries

Generative biology is full of huge claims. I wanted the opposite: a safe toy benchmark that borrows the shape of masked sequence modeling without pretending to design real proteins.

proteinmask builds synthetic protein-like families with conserved positions, trains a profile model, fills masks, and writes toy sequences to FASTA. The README says plainly that these are not wet-lab candidates.

Recover The Motif Or Stop Talking

The benchmark hides conserved positions in held-out sequences and checks whether the learned profile recovers them. That is the smallest honest version of the task.

The generator also tracks novelty so it cannot pass by copying the training strings. It has to preserve the motif while producing new toy sequences.

The Number

Profile motif recovery is 0.915. The random baseline is 0.000. Generated toy sequence novelty is 1.000.

3 tests pass locally and on GitHub Actions across Python 3.9, 3.11, and 3.13, with `artifacts/designs.fasta` regenerated by the benchmark.

Tools Used

Python
Masked Modeling
Sequence Design
Motif Recovery
FASTA
pytest
GitHub Actions CI