some advices about ggcoverage.R

[champeil]

When I read the track_df from bw files, I found the regions are not sorted, and the variants 'region.start' and 'region.end' are retrieved from the first line and the last line, all of which will cause some valid mark regions will be judged as invalid regions. And maybe the two variants can retrieved as 'min(start)' and 'max(end)', so it can make sure the mark regions will be determined truly. The suggests are from the geom_coverage.R code