chai1-structure-predictionlisted

# Chai-1 Structure Prediction **Plain-language role**: Use Chai when you want a modern structure-prediction model for validating designed binders or complexes. ## Prerequisites | Requirement | Minimum | Recommended | |-------------|---------|-------------| | Python | 3.10+ | 3.11 | | CUDA | 12.0+ | 12.1+ | | GPU VRAM | 24GB | 40GB (A100) | | RAM | 32GB | 64GB | ## How to run > **First time?** See [Installation Guide](../../docs/installation.md) to set up Modal and biomodals. ### Option 1: Modal ```bash cd biomodals modal run modal_chai1.py \ --input-faa complex.fasta \ --out-dir predictions/ ``` **GPU**: A100 (40GB) | **Timeout**: 30min default ### Option 2: Chai API (recommended) ```bash pip install chai_lab python -c " import chai_lab from chai_lab.chai1 import run_inference # Run prediction run_inference( fasta_file='complex.fasta', output_dir='predictions/', num_trunk_recycles=3 ) " ``` ### Option 3: Local installation ```bash git clone https://github.com/chaidiscovery/chai1-structure-prediction-lab.git cd chai1-structure-prediction-lab pip install -e . chai1-structure-prediction-lab predict \ --fasta complex.fasta \ --output predictions/ ``` ## FASTA Format ### Protein complex ``` >binder MKTAYIAKQRQISFVKSHFSRQLE... >target MVLSPADKTNVKAAWGKVGAHAGE... ``` ### Protein + ligand ``` >protein MKTAYIAKQRQISFVKSHFSRQLE... >ligand|smiles CCO ``` ### Protein + DNA/RNA ``` >protein MKTAYIAKQRQISFVKSHFSRQLE... >dna ATCGATCGATCG ``` ## Key param