Added na.rm parameter for getMeth regions.

[aerval]

If a given region contains CpGs with with coverage 0, they are currently excluded from the averaging of the methylation values in that region. This can be highly biased if CpGs in the region have different methylation values and are compared to samples with higher coverage. Now you get NA as return if na.rm = FALSE for those regions, so you can easily exclude them from analysis. In theory I would think that this might even be the default parameter, though I understand that we should not force it here for compatibility reasons.

Notes:

1. I am not sure whether the argument makes much sense for smoothed beta-values. I still included it from comparability
2. I have not that much experience in R-documentation, please check whether my documentation is correct (or do you use any framework like Roxygen for that?)

[codecov-io]

Current coverage is 16.68% (diff: 0.00%)

Merging #45 into master will not change coverage

@@             master        #45   diff @@
==========================================
  Files            18         18          
  Lines          2301       2301          
  Methods           0          0          
  Messages          0          0          
  Branches          0          0          
==========================================
  Hits            384        384          
  Misses         1917       1917          
  Partials          0          0          

Powered by Codecov. Last update 364f1fb...0e4c82a