[galaxy-user] line estimation for pileup generation

Hello, I am curious if the line estimation shown in the history window for pileup generation is at all accurate. I am using the pileup files to generate expression data from bwa mapping for looking at differential expression, but I am having some trouble understanding the line estimates. For example, for one pileup file, when I cut the reference id column and the number of hits column (columns 1 and 4), the number of lines in the cut file is about 25% that of the pileup file, and for another file it will be 5000%. How can the number of lines grow 50x when I am just cutting columns from the file? Shouldnt the line estimate be the same? Thanks, Austin