inst/scripts/andy2011.R

In lgatto/pRolocdata: Data accompanying the pRoloc package

library("MSnbase")
library("pRoloc")
library("pRolocdata")

###########################################################
## From Andy's protein file "ispy_results_proteins_stringent.csv"
## Peptides were merged into proteins by intensity weighted mean
andy <- read.csv("../extdata/andy2011_ispy_results_proteins_stringent.csv.gz",
                 row.names = 1)
.exprs <- as.matrix(andy[, 5:12])
.fData <- andy[, c(1:4, 14, 15, 13)]
.fData <- new("AnnotatedDataFrame",data = data.frame(as.matrix(.fData)))
.fData@varMetadata[,1] <- c("UniProtKB accession number",
                            "Full protein name",
                            "Peptides",
                            "Peptide spectrum match",
                            "PhenoDisco output as described in Breckels et al (2013) Journal of Proteomics. Accepted February 2013", 
                            "Updated protein markers (original markers see column 'markers' plus new protein markers found by phenoDisco and verified by UniprotKB/literature as described in Breckels et al)",
                            "Protein markers")
.pData <- new("AnnotatedDataFrame",
              data=data.frame(Fraction.information = c(rep("High Density", 7),
                                                       "Soluble/cytosolic"),
                              row.names = colnames(.exprs)))

.experiment <- new("MIAPE",
                   lab = "Cambridge Centre for Proteomics (CCP)",
                   name = "Andy Christoforou",
                   contact = "Kathryn S. Lilley",
                   email = "k.s.lilley@bioc.cam.ac.uk",
                   samples = list(
                       species = "Homo sapiens",
                       tissue = "Cell",
                       cellLine = "Embryonic Kidney Fibroblast Cells (HEK293T)",
                       operator = "Andy Christoforou"),
                   title="",
                   abstract = "",
                   pubMedIds = "",
                   url = "http://proteomics.bio.cam.ac.uk/",
                   instrumentModel = "LTQ Orbitrap Velos",
                   instrumentManufacturer = "ThermoScientific",
                   ionSource = "New Objective PicoView nano-electrospray",
                   analyser = "Orbitrap",
                   detectorType = "Orbitrap",
                   softwareName = "Mascot Search Engine",
                   isolationWidth = 0.5,
                   collisionEnergy = "45% CE HCD (two-stepped, 10% width)",
                   other = list(
                       experimentType = "Standard LOPIT experimental design on HEK293T. Label: iTRAQ 8-plex. Instrument: LTQ Orbitrap Velos. Scan Mode for Identification: MS2-HCD. Scan Mode for Quantitation: MS2-HCD. Scan Setup: Nth order double play, Top 10 HCD. MS1 Scan: FTMS, resolution = 30000, scan range m/z 380 - 1600. MS2 Scan: FTMS, resolution = 7500, scan range m/z 100 - 2000. Precursor Ion Selection Window: 0.5 Da (very stringent). Collision Energy: 45% CE HCD (two-stepped, 10% width)",
                       searchParameters = "Search Engine: Mascot. Search Database: UniProt Human. Fixed Modifications: iTRAQ 8-plex (N-term), iTRAQ-8plex (K), MMTS (C). Variable Modifications: iTRAQ 8-plex (Y), Oxidation (M). Enzyme: Trypsin. Max. Missed Cleavages: 2. Decoy Type: None (see Percolator parameters below). Peptide Charge: 5+. Peptide Tolerance: +/- 25 ppm. MS/MS Tolerance: +/- 0.2 Da. Instrument: ESI-ORBITRAP-HCD",
                       postProcessing = "Maximum Parent Proteins: 1. Unquantifiable spectra (E-value for PSM > 0.05, non-unique sequences, very low ion counts, >2 zero value reporter ions ) removed. Spectra filtered based on several criteria (precursor relative signal, position of switch relative to peak apex, reporter ion intensity) to pick a single peptidotypic spectrum per peptide. Peptides were merged into proteins by intensity weighted mean",
                       MS = "iTRAQ8",
                       spatexp = "LOPIT",
                       markers.fcol = "pd.markers",
                       prediction.fcol = "pd.2013"
                   ),
                   dateStamp = "2011-07")

.process <- new("MSnProcess",
                processing = c(
                    paste("Loaded on ", date(), ".",sep = ""),
                    paste("Normalised to sum of intensities.")),
                normalised = TRUE,
                files = "andy2011_ispy_results_proteins_stringent.csv")

andy2011 <- new("MSnSet",
                exprs = .exprs,
                phenoData = .pData,
                experimentData = .experiment,
                featureData = .fData)
andy2011@processingData <- .process

## Add updated markers
load("../extdata/markersHuman.rda")
fData(andy2011)$markers.orig <- fData(andy2011)$markers
fData(andy2011)$markers <- NULL
andy2011 <- addMarkers(andy2011, markers = mrk, verbose = FALSE)

## Add transfer learning results from Breckels et al 2015
load("../extdata/tl-res/human-tl.rda")
andy2011 <- minMarkers(andy2011, 13)
ind = which(fvarLabels(andy2011) == "markers13")
fvarLabels(andy2011)[ind] <- "markers.tl"
experimentData(andy2011)@other$knntl <- human.tl$knntl
experimentData(andy2011)@other$svmtl <- human.tl$svmtl

## Update fvarMetaData slots
fvarMetadata(andy2011)["markers", 1] <- "Updated and curated marker list following Breckels et al 2013"
fvarMetadata(andy2011)["markers.tl", 1] <- "Markers used for transfer learning classification in Breckels et al 2015"

stopifnot(pRolocdata:::valid.pRolocmetadata(pRolocmetadata(andy2011)))

## Using stable UniProt accession numbers as feature names
fData(andy2011)$UniProtKB.entry.name <- featureNames(andy2011)
featureNames(andy2011) <- fData(andy2011)$Accession.No.

if (validObject(andy2011)) {
    save(andy2011, file = "../../data/andy2011.RData",
         compress = "xz", compression_level = 9)
    ## rename HEK293T2011 (new in version 1.11.9)
    HEK293T2011 <- andy2011
    save(HEK293T2011, file = "../../data/HEK293T2011.RData",
         compress = "xz", compression_level = 9)
}