coverage2cytosine dies when --merge_CpG and --zero were specified

[FelixKrueger]

"I'm having some problems with coverage2cytosine.
It's terminating with exit code 25, and at the end of the output are the lines:

he strand of line 1 and line 2 were not + and -:

line1:  Methylated      2       -       0       0       CG CGA
line2:  Methylated      21      +       25      1       CG CGG

This is an analysis of spike in methylation controls and so the genome is very small.

[FelixKrueger]

The genome sequence was: TCGAGAGGC..., so the first CG positions were 2 and 3. Genome wide CpG reports may report the methylation state of a CG on the top strand starting at position 1, but not for the C on the reverse strand at position 2 because it can't determine a tri-nucleotide context. Thus we discard only the first line and read in a new one if the position of the first C was 1.

Now when the option --zero is specified however the first cytosines are reported as positions 1 and 2, which triggers the above correction behaviour, which I have now changed to take the 0-based coords into account. The resulting coverage file was also changed to now use 0-based genomic start and 1-based genomic end coordinates (zero-based, half-open), like used in the bedGraph file, like so:

Methylated      1       3       96.000000       24      1
Methylated      21      23      97.500000       39      1
Methylated      24      26      100.000000      41      0

Addressed in 71a6600 and f45f2c2.