Install and benchmark a number of

metaRboomics molecular formulas

install.packages("microbenchmark")
install.packages("plyr")

install.packages("rcdk")
install.packages("OrgMassSpecR")
install.packages("CHNOSZ")

BiocManager::install("Rdisop")
BiocManager::install("MetaboCoreUtils")

BiocManager::install("ChemmineR")
BiocManager::install("ChemmineOB")

BiocManager::install("enviPat")


## These require loading the packages explicitly
library(plyr)

library(CHNOSZ)
library(enviPat)
library(MetaboCoreUtils)
library(rcdk)
library(ChemmineR)
library(ChemmineOB)
library(OrgMassSpecR)
library(Rdisop)

## enviPat
data(isotopes)

benchmark <- microbenchmark::microbenchmark(
              MetaboCoreUtils = MetaboCoreUtils::calculateMass("C2H6O"),
              rcdk = rcdk::get.formula("C2H6O", charge = 0)@mass,
              Rdisop = Rdisop::getMolecule("C2H6O")$exactmass,
              ChemmineR = ChemmineR::exactMassOB(ChemmineR::smiles2sdf("CCO")),
              OrgMassSpecR = OrgMassSpecR::MonoisotopicMass(formula = OrgMassSpecR::ListFormula("C2H6O)"), charge = 0),

              CHNOSZ = CHNOSZ::mass("C2H6O"),
              enviPat = enviPat::isopattern(isotopes, "C2H6O", charge=FALSE, verbose=FALSE)[[1]][1,1]
            , times=1000L)


masses <- c(
  MetaboCoreUtils=MetaboCoreUtils::calculateMass("C2H6O"),
  rcdk=rcdk::get.formula("C2H6O", charge = 0)@mass,
  Rdisop=Rdisop::getMolecule("C2H6O")$exactmass,
  ChemmineR=ChemmineR::exactMassOB(ChemmineR::smiles2sdf("CCO")),
  OrgMassSpecR=OrgMassSpecR::MonoisotopicMass(formula = OrgMassSpecR::ListFormula("C2H6O)"), charge = 0),
  CHNOSZ=CHNOSZ::mass("C2H6O"),
  enviPat=enviPat::isopattern(isotopes, "C2H6O", charge=FALSE, verbose=FALSE)[[1]][1,1]
)


options(digits=10)

t(t(sort(masses)))

summary(benchmark)[order(summary(benchmark)[,"median"]) , ]

clipr::write_clip(as.data.frame(summary(benchmark)[order(summary(benchmark)[,"median"]) , ] ))

• install.packages("microbenchmark") to make it runable on vanilla R
• • BiocManager::install("enviPat") was missing
• smiles2sdf --> ChemmineR::smiles2sdf since it was not found before when the package was not loaded
• t(t(sort(masses))) to make it easier to read
• name benchmark entries
• OrgMassSpecR had an unfair advantage since the heavy lifting is converting the formula to its components. Added OrgMassSpecR::ListFormula that does that.

RESULTS:

install.packages("microbenchmark")
install.packages("plyr")

install.packages("rcdk")
install.packages("OrgMassSpecR")
install.packages("CHNOSZ")

BiocManager::install("Rdisop")
BiocManager::install("MetaboCoreUtils")

BiocManager::install("ChemmineR")
BiocManager::install("ChemmineOB")

BiocManager::install("enviPat")


## These require loading the packages explicitly
library(plyr)

library(CHNOSZ)
library(enviPat)
library(MetaboCoreUtils)
library(rcdk)
library(ChemmineR)
library(ChemmineOB)
library(OrgMassSpecR)
library(Rdisop)

data(isotopes)


benchmark <- microbenchmark::microbenchmark(
              MetaboCoreUtils = MetaboCoreUtils::calculateMass("C2H6O"),
              rcdk = rcdk::get.formula("C2H6O", charge = 0)@mass,
              Rdisop = Rdisop::getMolecule("C2H6O")$exactmass,
              ChemmineR = ChemmineR::exactMassOB(ChemmineR::smiles2sdf("CCO")),
              OrgMassSpecR = OrgMassSpecR::MonoisotopicMass(formula = OrgMassSpecR::ListFormula("C2H6O)"), charge = 0),

              CHNOSZ = CHNOSZ::mass("C2H6O"),
              enviPat = enviPat::isopattern(isotopes, "C2H6O", charge=FALSE, verbose=FALSE)[[1]][1,1]
            , times=1000L)


masses <- c(
  MetaboCoreUtils=MetaboCoreUtils::calculateMass("C2H6O"),
  rcdk=rcdk::get.formula("C2H6O", charge = 0)@mass,
  Rdisop=Rdisop::getMolecule("C2H6O")$exactmass,
  ChemmineR=ChemmineR::exactMassOB(ChemmineR::smiles2sdf("CCO")),
  OrgMassSpecR=OrgMassSpecR::MonoisotopicMass(formula = OrgMassSpecR::ListFormula("C2H6O)"), charge = 0),
  CHNOSZ=CHNOSZ::mass("C2H6O"),
  enviPat=enviPat::isopattern(isotopes, "C2H6O", charge=FALSE, verbose=FALSE)[[1]][1,1]
)



options(digits=10)

t(t(sort(masses)))


summary(benchmark)[order(summary(benchmark)[,"median"]) , ]


clipr::write_clip(as.data.frame(summary(benchmark)[order(summary(benchmark)[,"median"]) , ] ))

Cool, thanks ! If this continues like this, maybe that turns into a section in the etaRbolomics-book one day :-) Yours, Steffen

Yeah. I think what this needs is some testing of what happens if you throw them odd things they might not be expecting like HCOOH, C3H2S0, Na+Cl-, H3O+, +H2O-, Fe4+, Cp3UCl, 2H2O (isotope notation I guess would vary), D2O.

Updated code with some improvements to RCDK: https://gist.github.com/zachcp/ac410b900d67264ffb8061d2b8fa45bf