NASA's next flagship observatory, the Roman Space Telescope, is expected to be launched in the late-2020s and will help us see wider into the high-redshift Universe. In order to maximize its scientific capabilities and synergy with other telescopes, theoretical models and simulations play an important role in forecasting the magnitude, number density, and spatial distribution of expected sources. Furthermore, these physical models are also essential to the interpretation of their intrinsic properties and physical origins. In this virtual lecture, I will showcase the wide range of predictions and data products released with the latest papers in the Semi-analytic forecasts series. I will also highlight some of the galaxy formation physics that are expected to be constrained by upcoming observations. These predictions are made with the well-established Santa Cruz semi-analytic model (SAM) and have been shown to perform extremely well at reproducing a variety of observed constraints for galaxies and AGN observed in the past. I will also give a brief demonstration on how they have been used in the planning of high-z Webb galaxy surveys and how to gain access to these publicly available mock catalogues.