@datagrok/bionemo

BioNeMo

The BioNeMo package integrates advanced models for protein structure prediction and molecular docking. It features EsmFold, which predicts 3D protein structures from amino acid sequences, and DiffDock, which predicts molecular interactions with proteins.

Setup

To use BioNeMo, you need an NGC API key for either cloud or local access.

• Cloud access – Run EsmFold, DiffDock, and MolMIM without installation using your API key.
• Local setup – Install DiffDock or MolMIM on your system:
  • DiffDock Setup
  • MolMIM Setup

Make sure your API key is properly configured for your chosen setup.

EsmFold model

EsmFold predicts 3D structures of proteins based on their amino acid sequences.

To apply EsmFold to an entire column:

• Navigate to the top menu and select Bio > BioNemo > EsmFold.
• In the dialog:
  • Choose the dataframe: Select the dataframe containing your sequences.
  • Select the column with sequences: Choose the column that contains the amino acid sequences.
• Click OK to process the sequences.

EsmFold will generate 3D structure predictions for all sequences in the selected column and save them in a new column.

To use EsmFold for a single sequence:

• Select the specific sequence in your dataset.
• The EsmFold panel will appear, showing the 3D structure prediction for that sequence.

DiffDock model

DiffDock predicts molecular interactions with proteins by generating 3D poses of these interactions.

To apply DiffDock to an entire column:

• Navigate to the top menu and select Chem > BioNemo > DiffDock.
• In the dialog:
  • Choose the dataframe: Select the dataframe with your data.
  • Specify the column with ligands: Choose the column containing the ligands for docking.
  • Select the target: Choose the protein target for the interaction.
  • Set the number of poses: Specify how many poses DiffDock should generate for each ligand.
• Click OK to start the process.

DiffDock will generate multiple poses for each ligand, identify the best one, and add them along with confidence values to a new column.

To view generated poses:

• Click on a pose in the dataset. The Mol* viewer will open to display the selected pose.
• In the Mol* viewer, use the combo popup to see additional poses and their confidence levels. Select a pose to view it.

To use DiffDock for a single cell:

• Select the structure in your dataset. The DiffDock panel will appear.
• Specify the target and number of poses.
• The Mol* viewer will show the generated poses and their confidence values.