Add bpi method (issue #168)

o Add bpi,OnDiskMsnExp method. Parameter initial allows to specify whether the
  original data from the mzML file is returned, or whether the values are
  calculated.
o Modify tic,OnDiskMSnExp method to always return the initial TIC from the
  mzML's header, unless parameter initial = FALSE
o Add unit tests for both methods.

From: jotsetung <johannes.rainer@gmail.com>

git-svn-id: file:///home/git/hedgehog.fhcrc.org/bioconductor/trunk/madman/Rpacks/MSnbase@124448 bc3139a8-67e5-0310-9ffc-ced21a209358

NAMESPACE

@@ -197,6 +197,7 @@ exportMethods(updateObject,
               intensity,
               mz,
               tic,
+              bpi,
               ionCount,
               fromFile,
               centroided, "centroided<-",

R/AllGenerics.R

@@ -142,3 +142,5 @@ setGeneric("filterFile", function (object, ...) standardGeneric("filterFile"))
 setGeneric("filterAcquisitionNum", function (object, ...) standardGeneric("filterAcquisitionNum"))
 
 ## isolationWindow generic is in mzR
+
+setGeneric("bpi", function(object, ...) standardGeneric("bpi"))

R/methods-OnDiskMSnExp.R

@@ -252,33 +252,41 @@ setMethod("rtime", "OnDiskMSnExp",
 ## Get the total ion Current (from the featureData, thus it will not
 ## be recalculated).
 setMethod("tic", "OnDiskMSnExp",
-          function(object, BPPARAM = bpparam()) {
-    skipFun <- c("pickPeaks")
-    if (length(object@spectraProcessingQueue) > 0) {
-        recalc <- any(unlist(lapply(object@spectraProcessingQueue,
-                                    function(z) {
-            if (any(z@FUN == skipFun))
-                return(FALSE)
-            return(TRUE)
-        })))
-    } else {
-        ## No need to calculate the peaks count; we can use the
-        ## information from the feature data.
-        recalc <- FALSE
-    }
-    ## Also enforce re-calculation if we've got totIonCurrent of 0!
-    if (any(fData(object)$totIonCurrent == 0))
-        recalc <- TRUE
-    if (recalc) {
-        vals <- unlist(spectrapply(object,
-                                   FUN = tic,
-                                   BPPARAM = BPPARAM))
-    } else {
-        vals <- fData(object)$totIonCurrent
-    }
-    names(vals) <- featureNames(object)
-    return(vals)
-})
+          function(object, initial = TRUE, BPPARAM = bpparam()) {
+              if (initial) {
+                  vals <- fData(object)$totIonCurrent
+              } else {
+                  ## Calculate the value.
+                  vals <- unlist(spectrapply(object, FUN = tic,
+                                             BPPARAM = BPPARAM))
+              }
+              names(vals) <- featureNames(object)
+              return(vals)
+          })
+
+############################################################
+## bpi
+##
+## Get the base peak intensity from each spectrum. If initial = TRUE
+## the basePeakIntensity from the feaureData is returned, i.e. the value
+## extracted from the header of the raw mzML file. If initial = FALSE the
+## actual base peak intensity will be calculated for each spectrum by first
+## applying all eventual data manipulation operations.
+setMethod("bpi", "OnDiskMSnExp",
+          function(object, initial = TRUE, BPPARAM = bpparam()) {
+              if (initial) {
+                  vals <- fData(object)$basePeakIntensity
+              } else {
+                  ## Calculate the value.
+                  vals <- unlist(spectrapply(object,
+                                             FUN = function(z) {
+                                                 return(max(intensity(z)))
+                                             }, BPPARAM = BPPARAM))
+              }
+              names(vals) <- featureNames(object)
+              return(vals)
+          })
+
 
 ############################################################
 ## collisionEnergy

inst/scripts/torture.R

@@ -1,24 +1,24 @@
 library(MSnbase)
 library(msdata)
-f <- msdata::proteomics(full.names = TRUE, pattern = "TMT_Erwinia")
+f <- msdata::proteomics(full.names = TRUE)
 
 ## readMSData
 for (i in 1:10000) {
     if (i %% 200 == 0)
         cat(i, "\n")
-    res <- readMSData(f)
-    sp <- spectra(res)
+    res <- readMSData(f[1], verbose = FALSE, msLevel. = 1)
+    res <- readMSData(f[1], verbose = FALSE, msLevel. = 2)
     sp <- res[[12]]
+    sp <- spectra(res)
 }
 
 ## readMSData2
 for (i in 1:7500) {
     if (i %% 200 == 0)
         cat(i, "\n")
-    res <- readMSData2(f, verbose = FALSE)
-    sp <- spectra(res)
-    sp <- res[[12]]
-    res <- readMSData2(mzf, verbose = FALSE)
+    res <- readMSData2(f[1], verbose = FALSE)
+    sp <- res[[6]]
+    res <- readMSData2(f[2], verbose = FALSE)
     sp <- spectra(res)
-    sp <- res[[12]]
 }
+

man/OnDiskMSnExp-class.Rd

@@ -37,6 +37,7 @@
 \alias{scanIndex,OnDiskMSnExp-method}
 \alias{spectra,OnDiskMSnExp-method}
 \alias{tic,OnDiskMSnExp-method}
+\alias{bpi,OnDiskMSnExp-method}
 \alias{precScanNum,OnDiskMSnExp-method}
 \alias{featureNames<-,OnDiskMSnExp-method}
 \alias{featureNames<-,OnDiskMSnExp,ANY-method}
@@ -400,17 +401,39 @@ Class \code{"\linkS4class{Versioned}"}, by class "pSet", distance 5.
     }
 
     \item{tic}{
-      \code{tic(signature(object="OnDiskMSnExp"))}:
-      get the total ion current of each spectrum in the data
-      set. This information is extracted from the object's
-      \code{featureData} and represents the tic provided in the original
-      raw data files.
+      \code{tic(signature(object="OnDiskMSnExp"), initial = TRUE,
+      BPPARAM = bpparam())}:
+      get the total ion current (TIC) of each spectrum in the data
+      set. If \code{initial = TRUE}, the information is extracted from
+      the object's \code{featureData} and represents the tic provided in
+      the header of the original raw data files. For \code{initial =
+      FALSE}, the TIC is calculated from the actual intensity values in
+      each spectrum after applying all eventual data manipulation methods.
+
+      \code{BPPARAM} parameter: see \code{spectra} method above.
+
+      Returns a numeric vector with names corresponding to the spectrum
+      names.
+    }
+
+    \item{bpi}{
+      \code{bpi(signature(object="OnDiskMSnExp"), initial = TRUE,
+      BPPARAM = bpparam())}:
+      get the base peak intensity (BPI), i.e. the maximum intensity from
+      each spectrum in the data set. If \code{initial = TRUE}, the
+      information is extracted from the object's \code{featureData} and
+      represents the tic provided in the header of the original raw data
+      files. For \code{initial = FALSE}, the TIC is calculated from the
+      actual intensity values in each spectrum after applying all
+      eventual data manipulation methods.
+
+      \code{BPPARAM} parameter: see \code{spectra} method above.
 
       Returns a numeric vector with names corresponding to the spectrum
       names.
     }
 
-    \item{featureNames}{      
+    \item{featureNames}{
       \code{tic(signature(object="OnDiskMSnExp"))}: return a
       \code{character} of length \code{length(object)} containing the
       feature names. A replacement method is also available.
@@ -494,7 +517,7 @@ Class \code{"\linkS4class{Versioned}"}, by class "pSet", distance 5.
     \item{\code{as(from, "MSnExp")}}{Converts the \code{OnDiskMSnExp}
       object \code{from}, to an in-memory \code{MSnExp}. Also available
       as an S3 method \code{as.MSnExp()}.}
-    
+
   }
 }
 

tests/testthat/test_OnDiskMSnExp_other_methods.R

@@ -176,3 +176,34 @@ test_that("Test precursor* for OnDiskMSnExp", {
     names(pcn) <- names(pcn2) <- NULL
     expect_identical(pcn, pcn2)
 })
+
+############################################################
+## bpi
+test_that("bpi,OnDiskMSnExp", {
+    ## Get the "initial" one.
+    basepi <- bpi(ondisk)
+    expect_identical(unname(basepi), fData(ondisk)$basePeakIntensity)
+    expect_identical(names(basepi), featureNames(ondisk))
+    ## Calculate:
+    basepi <- bpi(ondisk, initial = FALSE)
+    expect_identical(names(basepi), featureNames(ondisk))
+    sp <- spectra(ondisk)
+    basepi_calc <- unlist(lapply(sp, FUN = function(z) max(intensity(z))))
+    expect_identical(basepi, basepi_calc)
+})
+
+
+############################################################
+## tic
+test_that("tic,OnDiskMSnExp", {
+    ## Get the "initial" one.
+    totic <- tic(ondisk)
+    expect_identical(unname(totic), fData(ondisk)$totIonCurrent)
+    expect_identical(names(totic), featureNames(ondisk))
+    ## Calculate:
+    totic <- tic(ondisk, initial = FALSE)
+    expect_identical(names(totic), featureNames(ondisk))
+    sp <- spectra(ondisk)
+    totic_calc <- unlist(lapply(sp, FUN = tic))
+    expect_identical(totic, totic_calc)
+})