Fig. 5

Change in SSE relative to SCONCE, as more cells are added to the consensus (mean) analysis. Number of cells, \(\kappa\), in the joint analysis is shown along the x-axis, and change in SSE (relative to SCONCE) is shown on the y-axis. Each line shows the mean change in SSE across cells, with error bars showing \(\pm 1\) standard deviation. Colors show cell ordering, where furthest denotes adding the least similar cells first (i.e., furthest distance as defined by the Euclidean distance on SCONCE profiles) and nearest denotes adding the most similar cells first. Across all datasets and cell orderings, SSE quickly drops as more cells are added, with adding the nearest cells (green line) showing the fastest improvement. However, for this ordering, the decrease in SSE levels off after approximately 10 cells are added, and then slightly rises as more cells are added, due to rare copy number events getting averaged out