Fig. 3

Scenarios for browser use. a Displaying a segment of the genome. While no data are stored in cache (blank blocks), only data within the queried region need to be fetched from the server (colored blocks) and are stored in cache for later use. b Shifting display window. Only the part not in cache needs to be fetched from the server (colored blocks) and merged in cache. c Zooming out. Existing cache data are used to recalculate new cache at a coarser granularity level, after which non-overlapping data are requested. d Zooming in. Because no cached data exist at a finer granularity level, all data within the queried region need to be fetched at that level