Extended Data Fig. 6: Differences in counts between neighboring groups after the non-linearity correction.

A linear ramp would show a flat line difference between neighboring groups. Residual slopes in differences between neighboring groups are indicative of detector effects that are not fully corrected (denoted by the dashed lines connecting the median of group pairs). Above we show the distributions in differences between neighboring groups for three detector rows where row 15 corresponds to the central location of the trace, and rows 14 and 16 are the neighboring detector rows. Distributions are colored by the number of groups that are usable before reaching our saturation threshold. Note that we follow the same methodology as¹¹ in flagging an entire column as saturated when the central row saturates.