./
    The root folder contains all analysis scripts that were executed
    in order to obtain the results described in the paper.
    Please ensure that your R installation provides all packages
    necessary to run the code (ROCR, randomForest, Cairo)

results/
    *-rocs.rda
        Cross-validated receiver operating characteristic estimates
        for the sequence-database (i.e. theoretical generalization
        error estimates)
        Loading creates a variable 'roc' with
            roc$pred: results of calling ROCR::prediction
            roc$perf: the corresponding ROCR performance object

    *-perf.rda
        ROCR performance objects obtained after applying the
        classifier emntioned in the file name to the dataset mentioned
        in the file name.

data/
    ipi.HUMAN.v3.36.ox.phos-masses.rda
        Theoretical masses for all potential singly/doubly/triply
        phosphorylated tryptic peptides extracted from the IPI HUMAN
        databaes (version 3.36). Includes oxidized cognates (i.e.
        Oxidation(M)).

    masses_phos_MS2_2-class_gt30_5.rda
        Observed precursor masses for phosphorylated peptides, sorted
        by their degree of phosphorylation (1-3 PTMs, with
        Oxidation(M) merged in).

    SusScrofa_24April09.masses.rda
        Theoretical masses for all potential singly/doubly/T3/T4
        iodinated tryptic Sus scrofa peptides extracted from the
        NCBInr database (downloaded on April, 24th, 2009).
        
    masses_KSU_DW_PIG_LTQ-Orbitrap_2-class_gt32.rda
        Observed MS/MS precursor masses sorted by their degree of
        iodination.

fig/
    All figures used in the publication.

src/
    Source code of all scripts used to generate the results presented
    in the publication. Note that the LICENSE only applies to the
    contents of this directory.

intermediate/
    All intermediate results, in particular all training datasets
    which are N=128000 subsamples of the files in the data/ directory.
    The contents of this diectory need to be downloaded separately.

    *-ts.rda
        A data.frame with one CV set in [[1]]...[[10]] each, and the
        set of labels in [[11]]

    *-cv.rda
        $classifier
            The trained classifer used to generate the results
            reported in the publication.
        $pred
            The ROCR prediction object generated from the CV results
            for the classifier given in $classifer
        $scatter
            The cross-validation confusion matrix for the classifier