Covalent Binding and Template-Based Generation

[yuvalrom-ligo]

First, I want to express my sincere appreciation for this incredible project. The work you've done is truly inspiring and opens up exciting new possibilities.

I am interested in understanding the current capabilities and potential future enhancements regarding two specific areas of molecule generation:

1. Covalent Binding Constraints: Is there a way to guide the model to generate molecules that form a covalent bond with a specific target protein/pocket?

2. Template-Based Protein Structure Generation: For generating a hit for a specific target, is it possible to use a template of the protein's 3D structure, as a starting point?

I understand that boltz-2 supports templates and specific binding behaviors. However, I am unclear whether these capabilities extend to enforcing a covalent binding constraint when generating molecules with synflownet. I'm also wondering if it's sufficient to constrain boltz-2 to achieve this, or if a post-processing step to filter out non-covalent trajectories would be necessary.

Thank you for your time and for this fantastic work.

Best regards,

Yuval