tagmPredict probreturn: predictin and joint

[lgatto]

@ococrook - I have two questions/comments regarding tagmPredict (which are now merged in the master branch) - here's an example:

> data(tan2009r1)
> p <- tagmTrain(tan2009r1)
> p
Object of class "MAPparams"
Algorithm: MAP 
> res <- tagmPredict(tan2009r1, p, probreturn = "joint")
> fvarLabels(res)
 [1] "FBgn"                  "Protein.ID"            "Flybase.Symbol"       
 [4] "AccessionNo"           "EntryName"             "AccessionNoAll"       
 [7] "EntryNameAll"          "No.peptide.IDs"        "Mascot.score"         
[10] "No.peptide.quantified" "PLSDA"                 "pd.2013"              
[13] "pd.markers"            "markers.orig"          "markers"              
[16] "markers.tl"            "tagm.allocation"       "tagm.probability"     
[19] "tagm.joint"            "tagm.outlier"         

Whether I use probreturn = "prediction" or "joint", tagm.allocation and tagm.probability are always generated, and tagm.joint only in the latter case. tagm.allocation is always expected, but I was expecting tagm.probability to be produced only when probreturn = "prediction".

Then

> summary(fData(res)$tagm.joint)
   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
     NA      NA      NA     NaN      NA      NA     888 

[ococrook]

1) I didn't think there was a situation where you would want the full data but not the final predictions so returning both for "joint" seemed more sensible 2) there are some stray grammar mistakes - these are corrected in my branch shall I pull request?

[lgatto]

Please do so, thanks.