Fisher precision

Fix a bug in bamToFastq and add a test for it

It was correctly identifying reads whose “paired” flag was set, but then
it was incorrectly skipping over the read-pair that follows such an
incorrect read.

Also added a test for this behavior.

fix #117 by re-ordering contingency table. 

Allow compilation with FreeBSD/clang.

Update intersect.rst

update the output from "-u" option.

Update the format on code block

Add human hg38 genome file

Update overview.rst

Correction: redistrubute --> redistribute

Created with the UCSC fetchChromSizes [1] tool:

    fetchChromSizes hg38 > genomes/human.hg38.genome

[1] http://hgdownload.cse.ucsc.edu/admin/exe/linux.x86_64/fetchChromSizes

fix -slop with -pct

Fisher's Exact Test