R Programming for Mass Spectrometry

The supplement

Web: https://books.wiley.com/titles/9781119872351/

Data download instructions

The supplement is a zip file with .html for each chapter but it is more useful to convert to markdown, and it is handy to run from a Linux terminal than otherwise nevertheless the markdown format renders can be rendered as follows,

for f in data-analysis intro-ms wrangle-data eda spectra-analysis chrom machine-learning
do
  export f=${f}
  pandoc ${f}.html --lua-filter=html.lua -t markdown -o ${f}.Rmd
  sed -i 's/``` {.r/```{r/' ${f}.Rmd
  Rscript -e 'f=Sys.getenv("f");rmarkdown::render(paste0(f,".Rmd"),output_dir="output"))'
done

with html.lua. For instance, in order to run through the R code,

  1. data-analysis.Rmd requires caution over Bash code blocks Rscript hello.R and R CMD BATCH hello.R, and a not-R code block.
  2. intro-ms.Rmd needs c("tidyverse").
  3. wrangle-data.Rmd requires c("tidyverse") and "tandem_result/" created by X!Tandem (tandem.sh, input.xml, default_input.xml, taxonomy.xml) shown here following ftp://ftp.thegpm.org/projects/tandem/source/.
  4. eda.Rmd requires c("Spectra").
  5. spectra-analysis.Rmd needs c("tidyverse", "Spectra", "infer", "xml2", "mzID", "MSnbase") as with inten_label and pal.
  6. chrom.Rmd needs c("tidyverse", "baseline", "signal", "EnvStats", "MassSpecWavelet", "MSnbase", "xcms", "latex2exp", "ggpubr", "fda.usc") as with inten_label and pal.
  7. machine-learning.Rmd requires c("tidymodels", "tidyverse", "visdat", "ggfortify", "factoextra", "colino", "heatmaply", "Spectra").

Set options(lifecycle_verbosity = "quiet") to use progress_estimated() in wrangle-data.Rmd, but a switch has been suggested 

library(progress)

n <- 100
pb <- progress_bar$new(
  format = "  processing [:bar] :percent eta: :eta",
  total = n, clear = FALSE, width = 60
)

for (i in seq_len(n)) {
  pb$tick()
  Sys.sleep(0.1)
}

inten_label and pal are from intro-ms.Rmd and data-analysis.Rmd, respectively. Batch load of packages can be done, e.g., pkgs <- c("tidyverse", "Spectra", "infer", "mzID", "MSnbase"); lapply(pkgs,library,character.only = TRUE).

large-data/mona/ (Chapter 7)

MoNA-export-LC-MS-MS_Positive_Mode.msp

MTBLS4938 (Chapter 7)

large-data/MSV000081318/MSV000086195

We start with wget

wget -r -nH --cut-dirs=2 -R "index.html*" ftp://massive-ftp.ucsd.edu/v01/MSV000081318/
wget -r -nH --cut-dirs=1 -R "index.html*" ftp://massive-ftp.ucsd.edu/v03/MSV000086195/

Directory listing including file transfer can also be done with

ftp massive-ftp.ucsd.edu <<EOF
anonymous
ls
cd z01/MSV000086195/ccms_peak/
prompt
mget *
EOF

where anonymous is the user name, or preferably by lftp,

lftp massive-ftp.ucsd.edu <<EOF
mirror --parallel=10 --verbose /v03/MSV000086195 ./MSV000086195
bye
EOF
# to resume
lftp -e "mirror --continue --parallel=4 /z01/MSV000086195/ccms_peak/ ccms_peak/; quit" \
      ftp://massive-ftp.ucsd.edu

ScltlMsclsMAvsCntr_Batch1_BRPhsFr5_prof.mzML in Chapters 4 & 5 is made with MSConvert (6GB!) or ThermoRawFileParser/1.4.4 (6.2GB with -p but 750MB without) as in mzML.sh following exercises in the Caprion project.

schema/ (Chapter 3):

Miscellaneous notes

This is a way around .mzid v1.2 (e.g., from i2MasChroQ 1.2.6) which neither PSMatch nor mzR supports; pyteomics has been made available from ~/rds/software/py3.11 therefore after source ~/rds/software/py3.11/bin/activate we have

from pyteomics import mzid
with mzid.read('ScltlMsclsMAvsCntr_Batch1_BRPhsFr29.mzid') as reader:
    first = next(reader)
    print(first['SpectrumIdentificationItem'])

[{'passThreshold': True, 'rank': 1, 'calculatedMassToCharge': 751.43574765, 'experimentalMassToCharge': 752.4484130691, 'chargeState': 2, 'PeptideEvidenceRef': [{'isDecoy': False, 'start': 30, 'end': 41, 'pre': 'L', 'post': 'F', 'PeptideSequence': 'ARLLVVYPWTQR', 'accession': 'sp|P11025|HBE_DIDVI', 'length': 147, 'Seq': 'MVHFTPEDKTNITSVWTKVDVEDVGGESLARLLVVYPWTQRFFDSFGNLSSASAVMGNPKVKAHGKKVLTSFGEGVKNMDNLKGTFAKLSELHCDKLHVDPENFRLLGNVLIIVLASRFGKEFTPEVQASWQKLVSGVSSALGHKYH', 'protein description': 'Hemoglobin subunit epsilon-M OS=Didelphis virginiana OX=9267 GN=HBE1 PE=2 SV=2', 'location': 'D:/Downloads/tandem_result/uniprot_sprot.fasta', 'FileFormat': 'FASTA format', 'DatabaseName': {'DatabaseName': 'uniprot_sprot.fasta'}, 'DB composition target+decoy': '', 'decoy DB accession regexp': '^XXX', 'decoy DB type reverse': ''}, {'isDecoy': False, 'start': 23, 'end': 34, 'pre': 'L', 'post': 'Y', 'PeptideSequence': 'ARLLVVYPWTQR', 'accession': 'sp|P02134|HBB_PELES', 'length': 140, 'Seq': 'GSDLVSGFWGKVDAHKIGGEALARLLVVYPWTQRYFTTFGNLGSADAICHNAKVLAHGEKVLAAIGEGLKHPENLKAHYAKLSEYHSNKLHVDPANFRLLGNVFITVLARHFQHEFTPELQHALEAHFCAVGDALAKAYH', 'protein description': 'Hemoglobin subunit beta OS=Pelophylax esculentus OX=8401 GN=HBB PE=1 SV=1', 'location': 'D:/Downloads/tandem_result/uniprot_sprot.fasta', 'FileFormat': 'FASTA format', 'DatabaseName': {'DatabaseName': 'uniprot_sprot.fasta'}, 'DB composition target+decoy': '', 'decoy DB accession regexp': '^XXX', 'decoy DB type reverse': ''}], 'X!Tandem:expect': 0.00173766, 'X!Tandem:hyperscore': 27.0, 'PeptideSequence': 'ARLLVVYPWTQR'}]

One could generate a CSV file as well through mzid.py.

Reference

Julian RK (2025). R Programming for Mass Spectrometry: Effective and Reproducible Data Analysis. ISBN: 978-1-119-87235-1.