do_findChromPeaks_centWave returns wrong maxo and int* values about xcms HOT 3 OPEN

I dove a little bit deeper into the code and it's likely an indexing issue. I've extracted the values that go into getEIC in order to extract the result from the function. Using the variables from the previous code gives us the following plot (black = original chromatogram, red = EIC chromatogram):

scanindex <- 1:length(rts) - 1
mzrange <- c(1, 1)
sr <- c(1, length(rts))

eic <- .Call("getEIC", mzs, intensities, as.integer(scanindex), as.double(mzrange),
             as.integer(sr), as.integer(length(scanindex)),
             PACKAGE = "xcms"
)

plot(intensities, type = "l", ylim = c(0, 2e5))
lines(eic$intensity, col = "red")

It seems that the getEIC function adds the previous scan intensity to the current scan intensity. A one-liner using dplyr confirms this:

points(intensities + dplyr::lag(intensities, n = 1), col = "green")

I haven't found the solution yet, but it's either in the loop in getScanEIC, or in the lowerBound or upperBound functions:

from xcms.

Hi, thanks for the report, we should turn this into a unit test once we've figured exactly what's going on.
Yours, Steffeb

from xcms.

Thanks Steffen,

Interestingly, this issue does not occur in the S4 method findChromPeaks. It specifically occurs in the do_* version of the method.

# Load library
library(xcms)

# Select MRM mzML from msdata package
data <- readSRMData(system.file("proteomics", "MRM-standmix-5.mzML.gz", package = "msdata"))

# Use 0.1 - 1 due to RT being in minutes
data <- findChromPeaks(data[5, ], CentWaveParam(peakwidth = c(0.1, 1)))
chromPeaks(data)

Returns the top peak with the correct maxo of 61141.1367

from xcms.