You best ask this question in the Rosetta Commons Forum:

https://www.rosettacommons.org/forum

That's where you are most likely to reach both experienced Rosetta users and developers. They are best able to tell you whether or not it is at all possible to take into account the different environments of the membrane, membrane surface and aqueous phase surrounding the different parts of a membrane proteins.

Thanks Annemarie Honegger !

I would not say that nobody does that, although most predictions address here multipass transmembrane proteins, and predictions of how multiple transmembrane helices interact in the context of the membrane.

Evolutionary-guided de novo structure prediction of self-associated transmembrane helical proteins with near-atomic accuracy

https://www.nature.com/articles/ncomms8196

RosettaMP is the Rosetta application you would use if you want to use Rosetta to predict Protein/Membrane systems:

https://www.rosettacommons.org/docs/latest/application_documentation/membrane_proteins/RosettaMP-GettingStarted-Overview

https://www.rosettacommons.org/docs/latest/application_documentation/membrane_proteins/RosettaMP-KeyElements-EnergyFunction

Alford RF, Koehler Leman J, Weitzner BD, Duran AM, Tilley DC, Elazar A, et al. (2015) An Integrated Framework Advancing Membrane Protein Modeling and Design. PLoS Comput Biol 11(9): e1004398. (*equal contribution authors)

Alford RF, Fleming PJ, Fleming KG, Gray JJ (2019) "Protein structure prediction and design in a biologically-realistic implicit membrane" BioRxiv

Koehler Leman J, Mueller BK, Gray JJ (2018) "Expanding the toolkit for membrane protein modeling in Rosetta" Bioinformatics 33 (5), 754-756

You might want to look at this paper for a starting point:

Article A lipophilicity-based energy function for membrane-protein m...

and at publications that reference this paper

You have to take into account that the experimental structural database for full-length single-pass transmembrane proteins is very small - usually these do not crystallize. Therefore, intracellular domains and extracellular domains are crystallized separately, and structures of the transmembrane domains are usually NMR structures of the helix in detergent mycelles.