Single-strand mode is not appropriate with unmerged R2 reads

[stschiff]

We realised that the current damage model used in --singleStrandMode is not appropriate if the input pileup data contains alleles from unmerged R2 reads from paired-end sequencing. Currently, this means, that users should make sure their input data does not contain unmerged reads, should they use data from paired-end sequencing.

In the longer run, I need to come up with a solution that recognises R2 reads in the pileup and treats them as reverse complements. This is not possible with the pileup-format produced by samtools mpileup, which is currently the only input format that pileupCaller accepts.

[stschiff]

I have for now added a line to the README, and might pick this up in the future