Dear BoltzGen Team,
First of all, thank you for developing such an impressive platform for molecular generation and screening. The integration of generative design, folding, and affinity evaluation is extremely valuable for researchers working on AI-driven drug discovery. We particularly appreciate the effort your team has put into building both the BoltzLab web interface and the local toolkit—these tools significantly lower the barrier for applying modern generative models in practical molecular design workflows.
We have successfully used the BoltzLab web platform to perform small-molecule binder generation and evaluation tasks, and the results have been very encouraging. The workflow is intuitive and the generated candidates are highly interesting from a drug design perspective.
Recently, we also deployed the local version of BoltzGen, as we are hoping to run small-molecule AI drug design workflows locally and explore more advanced or customizable configurations. We are very excited about the possibility of using the local pipeline to perform target-driven molecular design.
However, based on the examples currently available in the GitHub repository, we were not able to determine the correct command-line configuration or YAML specification needed to generate small-molecule binders for a given target protein. Most examples appear to focus on protein or peptide design, and we were unable to construct a YAML file that successfully triggers small-molecule binder generation.
Therefore, we would like to ask whether the current local version of BoltzGen supports generating small-molecule binders directly from a given target protein structure.
Specifically, we would like to know:
Are there specific configuration flags, constraints, or input formats required to enable small-molecule generation?
Is there a recommended way to restrict the generation to small molecules (for example via molecular weight constraints)?
Do we need to specify inputs such as SMILES strings, ligand templates, or a particular “ligand mode” in the YAML design specification?
If there are any example YAML files, documentation, or recommended configurations for this type of task, we would greatly appreciate it if you could point us to them.
Thank you again for creating such an exciting platform. We look forward to exploring more capabilities of BoltzGen and would be very grateful for any guidance you can provide.
Best regards,
Dear BoltzGen Team,
First of all, thank you for developing such an impressive platform for molecular generation and screening. The integration of generative design, folding, and affinity evaluation is extremely valuable for researchers working on AI-driven drug discovery. We particularly appreciate the effort your team has put into building both the BoltzLab web interface and the local toolkit—these tools significantly lower the barrier for applying modern generative models in practical molecular design workflows.
We have successfully used the BoltzLab web platform to perform small-molecule binder generation and evaluation tasks, and the results have been very encouraging. The workflow is intuitive and the generated candidates are highly interesting from a drug design perspective.
Recently, we also deployed the local version of BoltzGen, as we are hoping to run small-molecule AI drug design workflows locally and explore more advanced or customizable configurations. We are very excited about the possibility of using the local pipeline to perform target-driven molecular design.
However, based on the examples currently available in the GitHub repository, we were not able to determine the correct command-line configuration or YAML specification needed to generate small-molecule binders for a given target protein. Most examples appear to focus on protein or peptide design, and we were unable to construct a YAML file that successfully triggers small-molecule binder generation.
Therefore, we would like to ask whether the current local version of BoltzGen supports generating small-molecule binders directly from a given target protein structure.
Specifically, we would like to know:
Are there specific configuration flags, constraints, or input formats required to enable small-molecule generation?
Is there a recommended way to restrict the generation to small molecules (for example via molecular weight constraints)?
Do we need to specify inputs such as SMILES strings, ligand templates, or a particular “ligand mode” in the YAML design specification?
If there are any example YAML files, documentation, or recommended configurations for this type of task, we would greatly appreciate it if you could point us to them.
Thank you again for creating such an exciting platform. We look forward to exploring more capabilities of BoltzGen and would be very grateful for any guidance you can provide.
Best regards,