Quick Start

Get up and running with BioDreamer in minutes.

Installation

pip

pip install bio-dreamer

Requires Python ≥ 3.9 and PyTorch ≥ 2.1

Basic Usage

from biodreamer import ProteinDreamer

# Load a pre-trained world model
dreamer = ProteinDreamer.from_pretrained("biodreamer/protein-dreamer-v1")

# Define wild-type sequence and objective
result = dreamer.design(
    sequence="MVLSPADKTNVKAAWGKVGAHAGEYGAEALERMFLSFPTTKTYFPHFDLSH...",
    objective="stability",
    n_steps=10,
    n_candidates=50,
)

# Ranked mutation paths dreamed in the world model
for candidate in result.top_candidates(5):
    print( candidate.mutations, candidate.predicted_fitness)

What's Next

This section will expand as the project matures. Full documentation will cover advanced configuration, custom training, active learning loops, and the API server.

Explore Features How It Works Full Documentation (coming soon)