From Genome Analysis Wiki
*Assumes the file is sorted by Coordinate (or ReadName if using <code>--readName</code> option)
*Assumes only 2 reads have matching ReadNames
**It matches in pairs, so if there are 3, the first 2 will be matched and compared, but the 3rd won't. If there are 4, the first 2 will be matched and the last 2 will be matched and compared.
*Only mapped reads will be clipped