I recently submitted an order to SPRING for two fusion proteins. These two fusion proteins should interact at with each other at two different sites. The SPRING results looked very clean and the models looked as would be expected on PDB. However, the generated models only provided the protein complex threading results for one of the two interaction sites between the two molecules (ie basically half of each submitted sequence was omitted in the modeling results). When I submitted an order to TACOS to model the full domains of each molecule, the nice protein interaction modeling obtained from the SPRING results completely disappeared. (Although there was one TACOS model which came somewhat close to modeling the second protein-protein interaction site).
Am I missing something here? Should the threading results and protein complexes found in SPRING be similar to the results for TACOS? Or is there some way for my order to include this constraint in TACOS (or PDB template, like with I-TASSER)?
Thank you very much for your time and effort!