TODO
-----------------------

    o remove naming of ORFs with "_"
    o remove unnecesary type checking
    o make computeFeatures completely independent on CageSeq and as input txdb
    o split files into smaller files, tests too
    o remove whole utils file?

    CageSeq:
    o make CageSeq leader reassignment return txdb object and take in txdb
    object
    o remove cageFromFile etc. so that we don't load the data for the user
    o remove filterCage so that we don't filter, user filters himself
    o matchSeqnames should be generalized to matchSeqlevels(grA, grB) etc.,
    I think it already exists as ?relevel

    RiboSeq:
    o add examples to getStartStopWindows
    o parseCigar implement into C++

    Features:
    o better docs with inherit params
