applyReFraction: Main function of Re-Fraction

In PengyiYang/ReFraction: A machine learning approach for deterministic identification of protein homologues and splice variants in large-scale MS-based proteomics

Description

This is the main function that apply Re-Fraction algorithm for deterministic protein identifications.

Usage

ReFraction(protein.database, peptide.table)

Arguments

protein.database

The protein database that was used for peptide identification. This database should have been processed to extract protein properties including: protein id, protein mass, protein length, number theoretic tryptic peptides, and protein pI. See the supplied example database for detail. There is a function for extracting protein properties from fasta file. See the example at the bottom.

peptide.table

ReFraction accepts peptide table that including fraction information and protein assignments. For fraction information, peptide table should contain columns with names start with "Slice" end with a number and separated by ".". An example of such a column would be "Slice.1". For protein assignments, there should be a column with name "Proteins" and contains one or more protein identifiers that the peptide is assigned to. The identifier should be separated by ";". Column names are case sensitive. The detailed structure of such a peptide table could be found in the example dataset. The detail of this dataset could be found in refernce [2].

Value

determine.proteins

This is a list of deterministic protein identifications after applying Re-Fraction. Proteins in the list are ranked by their posterior error probability (PEP). The smaller the PEP the more confident the protein identified.

Author(s)

Pengyi Yang

References

[1] Pengyi Yang, Sean J. Humphrey, Daniel J. Fazakerley, Matthew J. Prior, Guang Yang, David E. James, and Jean Yee-Hwa Yang, Re-Fraction: a machine learning approach for deterministic identification of protein homologs and splice variants in large-scale MS-based proteomics, Journal of Proteome Research, http://dx.doi.org/10.1021/pr300072j

[2] Matthew J. Prior, Mark Larance, Robert T. Lawrence, Jamie Soul, Sean Humphrey, James Burchfield, Carol Kistler, Jonathon R. Davey, Penelope J. La-Borde, Michael Buckley, Hiroshi Kanazawa, Robert G. Parton, Michael Guilhaus, and David E. James, Quantitative proteomic analysis of the adipocyte plasma membrane, Journal of Proteome Research, 2011, 10(11), 4970-4982

Examples

# load the example protein database that used for peptide identification
data("ipiMOUSEv385")
# load the example peptide table
data("peptideTable")

# to have a view of how extracted protein database should look like
protein.database[1:10,]

# to have a view of how peptide table should look like. There are extra columns than what are required by Re-Fraction. The required columns are: "id", "Proteins", "PEP", and all the columns start wtih "Slice".
peptide.table[1:10,]

# perform Re-Fraction algorithm for deterministic protein identifications
ReFraction.results <- applyReFraction(protein.database, peptide.table, fdr.cutoff=0.01)

# display deterministic protein identifications; assigned peptide ids match those in input peptide.table
ReFraction.results$deterministic.protein.table[1:15,]

# display peptide table; the last column contains the refined peptide assignment after applying ReFraction
ReFraction.results$peptide.ReFraction.table[1:15,]

# The above is an example run on the dataset used in reference [1].
# For processing your dataset, you need to firstly extracting protien properties from the database you used in peptide identification.
# Use the following function for this:
?extractDatabase

PengyiYang/ReFraction documentation built on May 14, 2019, 11:01 p.m.