Initial guess for multimer structure prediction

[Rajat-Punia]

Is it possible to provide the initial guess for the multimer structure inference somehow?

I understand that AF3 use monomer structure as templates with an option to provide external monomer template, but is it possible to somehow provide the information on the interface contact residues e.g. in distogrm OR initial guess to the diffusion.

My question pertains to something similar done in past with AF2 (https://www.nature.com/articles/s41467-023-38328-5).

I have cases where AF3 fails to correct predict the multimer structure but believe that provide some partial additional information might improve accuracy.