Add methods for combining duplicate samples

[katehoffshutta]

A user may wish to combine information from duplicate samples, rather than dropping duplicates (which is the current supported functionality). We plan to add support for:

1. combining multiple RNA-seq duplicate samples by adding counts for each gene
2. combining multiple DNA methylation samples by averaging methylation values at each CpG probe site

[katehoffshutta]

We decided not to support this functionality because combining information from duplicates may mask important characteristics (e.g., contamination of a sample, batch effects, different cell type composition) and cause problems with downstream analysis.

We have, however, added two new ways to filter duplicates - filterDuplicatesRandom, which selects one duplicate to keep at random, and filterDuplicatesMethylationMissingness, which handles duplicate methylation samples by selecting the one with the least missing data. See #73 and #74 for details.