*3.5. Heterogeneity of Data*

The type and frequency of data gathered varied greatly among the studies reviewed. In some papers, only one figure with one retinal section was o ffered as evidence for PR degeneration, while other papers showed extensive quantitation of their data. To document potential sources of variability in the data, we indicate the method by which the degree of degeneration was determined, either by measuring ONL thickness or by count of nuclei in the ONL, typically the number of rows of nuclei spanning the ONL but sometimes a total count of ONL nuclei in a fixed area of a retinal micrograph. In some instances, when data was quantified in spider plots or bar graphs, mean values obtained from the central retina were used in estimating the PR loss. We normalized the data among studies by recording the percent degeneration as determined by dividing the mutant values by the corresponding values from age-matched controls as reported in each publication.

### *3.6. Comparison of Progressive PR Cell Loss*

To compare progressive PR cell loss among models, we fit normalized data from each article to an exponential decay that includes a delay, or o ffset [36]. Ranges of either age or photoreceptor numbers, if reported, were averaged. Fitting was performed in Excel Visual Basic using a piecewise equation that modeled the delay with a straight line at 100% and the remaining points with a monoexponential decay to 0%. Two adjustable parameters, the delay and the decay rate constant, were optimized. We calculated the age at which PR cell numbers reached 50% of control values (D50) as a measure of progression. Roughly one third of the datasets contained only a single point within the exponential regime, which was insu fficient to calculate D50. In these cases, D50 was calculated at the extremes of zero delay and infinite rate, and the mean of these values was used as a D50 estimate.

### *3.7. Generation of Primary Data Using Fundus Imaging and OCT Scans*

Fundus photographs of EMR mutants were taken in unanesthetized mice treated with 1% cyclopentolate to dilate or enlarge the pupil with an in vivo bright field retinal imaging microscope equipped with image-guided OCT capabilities (Micron III; Phoenix Laboratories, Inc., Pleasanton, CA, USA) as previously described [20]. This system allows for the visualization of the location of the OCT scan using the real-time Micron III bright-field image. A superimposed line placed directly on the image over the retinal feature being examined delivers precise cross-sectional information, allowing for the assessment of changes in layer thickness and morphological alterations.

Fundus photodocumentation for TVRM mutants and C57BL/6J control mice was performed using a Micron III or IV retinal camera (Phoenix Laboratories, Inc., Pleasanton, CA, USA) as described [37], except that 1% cyclopentolate or 1% atropine was used as a dilating agent, and in some cases, mice were anesthetized with isoflurane. OCT imaging to assess retinal layer thickness in *Nmnat1tvrm113*, *Ctnna1Tvrm5*, and C57BL/6J control mice was performed using a Bioptigen ultrahigh-resolution (UHR) Envisu R2210 spectral domain OCT (SDOCT) imaging system for volume scanning as described [37,38] with ketamine/xylazine (1.6 mL ketamine (100 mg/mL), 1.6 mL xylazine (20 mg/mL), and 6.8 mL sodium chloride (0.9% *w*/*v*)) as an anesthetic. A representative B-scan through the optic nerve head was derived from the OCT volume dataset. *Rpgrip1nmf247* and *Alms1Gt(XH152)Byg* were assessed on the same OCT system by obtaining a linear B-scan with the following parameters: length, 1.9 mm; width, 1.9 mm; angle, 0 degrees; horizontal o ffset, 0 mm; vertical o ffset, 0 mm; A-scans/B-scan, 1000 lines; B-scans, 1 line; frames/B-scan, 20 frames; and inactive A-scans/B-scan, 80 lines. Linear scans were registered and averaged in the InVivoVue program to merge the 20 frames into a single image.
