We approximated the amount of the ROS disk membrane accounted for by the SMALPs of our three chosen samples to assess the completeness of our analysis. We began with the fact that each SMALP will have a certain number of lipids copurifying with it. When ROS membranes are incubated with SMA, ∼12-nm-diameter disks of lipids encapsulate bilayer patches of the ROS, with each side of the bilayer representing roughly 113 nm² (equaling 226 nm² for the full surface area of each native nanodisc; Fig. 1 a). Because of the high density of proteins in the ROS disk membranes, we assumed protein-free SMALP-solubilized membranes would be negligible in amount. Then, with the ROS membrane separated into membrane protein–SMALPs, we estimate the number of lipids that can be accommodated within the SMALP in the presence of each of the known ROS membrane proteins, based the number of transmembrane helices that would occupy a portion of the total SMALP area (T_n). This is done using twice the average cross-sectional area of a transmembrane helix (1.4 nm²) to account for the helix displacing lipids in both bilayers (Eskandari et al., 1998; Swainsbury et al., 2014; Takamori et al., 2006). The transmembrane helix surface area is then subtracted from the total surface area of the nanodisc membrane to yield the available surface area for lipids within the SMALP. That surface area, divided by the average cross-sectional area of a single phospholipid (∼0.78 nm²), gives the number of lipids that should fit in the given ROS protein SMALP (Lee, 2003). The number of lipids that each membrane protein–SMALP can carry was then scaled by each protein’s relative abundance in the ROS membranes. The relative abundance, A_n, was calculated by cross-referencing the nine membrane proteins classified as ROS disk–specific with the ROS disk proteomics reported using absolute protein expression (APEX) measurements taken by tandem mass spectrometry (MS/MS; Skiba et al., 2013; Kwok et al., 2008). This is the estimation of what we call the weighted lipid contribution (WLC) of each protein (Eq. 1).WLC_rhodopsin, WLC_ABCA4, and WLC_PRPH2/ROM1 were added together and divided by the sum of all WLCs, giving an approximate lipid contribution of 95% from the three samples studied here (Eq. 2 and Table 2).This estimation gives us confidence that we have studied the majority of the ROS disk membranes.