R/ParseMoleculeInformation.R

In IsoCorrectoR: Correction for natural isotope abundance and tracer purity in MS and MS/MS data from stable isotope labeling experiments

# Analysis of molecule formula with regex

# The strings imported from MoleculeFile containing information on the number 
# of atoms per element are analyzed using regular expressions.  The regular
# expression Characters extracts all characters, the regular expression 
# Numbers extracts all numbers.  TracerElement looks for the tracer element
# identifier.

ParseMoleculeInformation <- function(MoleculeInformation, ElementInfo, UltraHighRes, verbose) {
  
  Characters <- "([A-Za-z]*)"
  Numbers <- "([0-9]+[0-9]|[0-9])"
  TracerElement <- "(Lab)"
  MoleculeFormulaRegex <- "(|Lab)[A-Z][a-z]?[1-9][0-9]*"
    
  if (is.na(MoleculeInformation[1,3])) {
    NumberFragments <- 1
  } else {
    NumberFragments <- 2
  }
  
  # collect features for each individual fragment
  tmpFragmentList <- list()
  
  # FragmentList contains one or two tmpFragmentList()s
  FragmentList <- list()
  
  # The following is done for each Molecule and each fragment 
  # of a given molecule
  for (Fragment in seq_len(NumberFragments)) {
    
    # For each molecule(-fragment), the string containing the element count 
    # information is split into the individual elements using a regex
    
    Elements <- stringr::str_extract_all(MoleculeInformation[1,Fragment + 1], MoleculeFormulaRegex)[[1]]
    
    # Now each element in Elements, consisting of the element ID and the 
    # element count, is again separated into ID and element count using 
    # the regular expressions constructed above. 
    # Additionally, a check is made to identify the 
    # tracer element information ('Lab').
    
    ElementCount.vec <- vector()
    Element.vec <- vector()
    TracerCount.vec <- vector()
    Tracer.vec <- vector()
    
    NumberElementsandTracer <- length(Elements)
    
    for (Element in seq_len(NumberElementsandTracer)) {
      
      # TRUE => current Element is Tracer // FALSE => current Element is no Tracer
      if (Elements[Element] %>% stringr::str_detect(TracerElement)) {
        
        TracerCount.vec <- c(TracerCount.vec, Elements[Element] %>% stringr::str_replace(TracerElement, "") %>% stringr::str_extract(Numbers) %>% 
                               as.numeric)
        Tracer.vec <- c(Tracer.vec, Elements[Element] %>% stringr::str_replace(TracerElement, "") %>% stringr::str_extract(Characters))
      } else {
        
        ElementCount.vec <- c(ElementCount.vec, Elements[Element] %>% stringr::str_extract(Numbers) %>% as.numeric)
        Element.vec <- c(Element.vec, Elements[Element] %>% stringr::str_extract(Characters))
      }  #else
    }  #Element    
    names(ElementCount.vec) <- Element.vec
    names(TracerCount.vec) <- Tracer.vec
    
    tmpFragmentList <- list(Element = ElementCount.vec, Tracer = TracerCount.vec)
    
    NumberElements <- length(ElementCount.vec)
    NumberTracers <- length(TracerCount.vec)
    
    # 'NumberElements' - 'NumberTracers' gives the number of non-tracer elements NumberElementsNonTracer.
    NumberElementsNonTracer <- NumberElements - NumberTracers
    
    if (!UltraHighRes) {
      
      # If a tracer is present in the molecule(-fragment) considered, 
      # the tracer parameters are extracted.  'MaxLabel' is the maximum amount
      # of tracer isotope that is expected to be found in the 
      # molecule(-fragment) due to metabolism while 'nTracerMax' is 
      # the maximum amount of tracer element (labelled or
      # unlabelled) in that same species. 'IDTracer' is the tracer elements ID.
      
      if (NumberTracers > 0) {
        
        MaxLabel <- max(TracerCount.vec)
        IDTracer <- TracerCount.vec %>% which.max %>% names
        
        # This part deals with the possibility of having a tracer element 
        # that shows isotopes with a negative mass shift.  
        # In this case IsoCombinationsMaster()
        # has to calculate `PlacesToAssign` differently for the tracer.
        
        if (sum(ElementInfo[[IDTracer]][["Isotopes"]][[1]][["MassShift"]] < 0) > 0) {
          NegIsoTracer <- 1  # => TRUE
        } else {
          NegIsoTracer <- 2  # => FALSE
        }
        
        names(NegIsoTracer) <- IDTracer
        
        tmp.which <- which(names(ElementCount.vec) %in% names(TracerCount.vec))
        nTracerMax <- as.numeric(ElementCount.vec[tmp.which])
        
        if (length(nTracerMax) == NumberTracers) {
          
          names(nTracerMax) <- names(TracerCount.vec)
          
        }
        
        names(MaxLabel) <- IDTracer
        
        tmpFragmentList[["MaxLabel"]] <- MaxLabel
        tmpFragmentList[["IDTracer"]] <- IDTracer
        tmpFragmentList[["nTracerMax"]] <- nTracerMax
        tmpFragmentList[["NegIsoTracer"]] <- NegIsoTracer
      } else {
        tmpFragmentList[["MaxLabel"]] <- NA
        tmpFragmentList[["IDTracer"]] <- NA
        tmpFragmentList[["nTracerMax"]] <- NA
        tmpFragmentList[["NegIsoTracer"]] <- NA
      }  #if(NumberTracers>0)
      
      # In this section, the non-tracer element parameters are extracted in 
      # the same way as for the tracer element.
      
      ElementsNonTracerList <- list()
      
      if (NumberElementsNonTracer > 0) {
        
        NonTracer <- 1
        
        tmpElementNonTracer.vec <- vector()
        tmpElementNonTracerCount.vec <- vector()
        tmpElementZeroTracer.vec <- vector()
        tmpElementZeroTracerCount.vec <- vector()
        
        for (Element in seq_len(NumberElements)) {
          # tracers exist
          if (NumberTracers > 0) {
            # current element is no tracer
            if (names(ElementCount.vec)[Element] != IDTracer) {
              tmpElementNonTracer.vec <- c(tmpElementNonTracer.vec, names(ElementCount.vec)[Element])
              tmpElementNonTracerCount.vec <- c(tmpElementNonTracerCount.vec, as.numeric(ElementCount.vec[Element]))
            }
          } else {
            tmpElementZeroTracer.vec <- c(tmpElementZeroTracer.vec, names(ElementCount.vec)[Element])
            tmpElementZeroTracerCount.vec <- c(tmpElementZeroTracerCount.vec, as.numeric(ElementCount.vec[[Element]]))
            
          }  #NumberTracers
        }  #Element
        
        names(tmpElementNonTracerCount.vec) <- tmpElementNonTracer.vec
        names(tmpElementZeroTracerCount.vec) <- tmpElementZeroTracer.vec
        
        tmpFragmentList[["NonTracer"]] <- tmpElementNonTracerCount.vec
        tmpFragmentList[["ZeroTracer"]] <- tmpElementZeroTracerCount.vec
        
      } else {
        # NumberElementsNonTracer>0
        
        tmpFragmentList[["NonTracer"]] <- c()
        tmpFragmentList[["ZeroTracer"]] <- c()
        
      }
      
    } else if (UltraHighRes) 
    {
      
      if (NumberTracers > 0) {
        
        # For multiple tracer correction, the gathering of tracer parameters 
        # has to loop through the number of tracers present in a 
        # given molecule (-fragment).
        
        MaxLabel.vec <- vector()
        IDTracer.vec <- vector()
        nTracerMax.vec <- vector()
        
        for (TracerNo in seq_len(NumberTracers)) {
          MaxLabel <- as.numeric(TracerCount.vec[TracerNo])
          IDTracer <- names(TracerCount.vec)[TracerNo]
          
          MaxLabel.vec <- c(MaxLabel.vec, MaxLabel)
          IDTracer.vec <- c(IDTracer.vec, IDTracer)
          
          for (Element in seq_len(NumberElements)) {
            # if current element is a tracer
            if (names(tmpFragmentList[["Element"]][Element]) == IDTracer) 
            {
              nTracerMax <- tmpFragmentList[["Element"]][Element] %>% as.numeric
              nTracerMax.vec <- c(nTracerMax.vec, nTracerMax)
            }  #if
          }  #Element
        }  #TracerNo
        names(MaxLabel.vec) <- IDTracer.vec
        
        if (length(nTracerMax.vec) == NumberTracers) {
          
          names(nTracerMax.vec) <- IDTracer.vec
          
        }
        
        tmpFragmentList[["MaxLabel"]] <- MaxLabel.vec
        tmpFragmentList[["IDTracer"]] <- IDTracer.vec
        tmpFragmentList[["nTracerMax"]] <- nTracerMax.vec
        
        # Gaining natural abundance information associated with the 
        # tracer isotopes for probability calculations
        
        # Number of isotopes per Element
        NumberIso <- unlist(lapply(ElementInfo, function(x) nrow(data.frame(x[["Isotopes"]]))))
        
        ### store information for each individual Tracer Element
        NatAbuTracerList <- list()
        NatAbuBaseList <- list()
        MassShiftTracerList <- list()
        
        MassShiftTracer.vec <- vector()
        
        for (TracerNo in names(TracerCount.vec)) {
          MassShiftTracer <- ElementInfo[[TracerNo]][[2]]
          
          # for each Tracer isotope
          NatAbuTracer.vec <- vector()
          NatAbuBase.vec <- vector()
          
          for (IsotopeNo in seq_len(NumberIso[TracerNo])) {
            
            if (data.frame(ElementInfo[[TracerNo]][[1]])[IsotopeNo, 2] == MassShiftTracer) {
              
              NatAbuTracer.vec <- c(NatAbuTracer.vec, data.frame(ElementInfo[[TracerNo]][[1]])[IsotopeNo, 1])  # 1=>IsotopeAbundance
              
            } else if (data.frame(ElementInfo[[TracerNo]][[1]])[IsotopeNo, 2] == 0) {
              
              NatAbuBase.vec <- c(NatAbuBase.vec, data.frame(ElementInfo[[TracerNo]][[1]])[IsotopeNo, 1])
              
            } else {
              
              if(verbose){message(date(), " :: skipping IsotopeNo ", IsotopeNo, " for Tracer ", TracerNo)}
              
            }
          }  #IsotopeNo
          
          NatAbuTracerList[[TracerNo]] <- NatAbuTracer.vec
          NatAbuBaseList[[TracerNo]] <- NatAbuBase.vec
          MassShiftTracerList[[TracerNo]] <- MassShiftTracer
          
        }  #TracerNo
        
        tmpFragmentList[["NatAbuTracer"]] <- unlist(NatAbuTracerList)
        tmpFragmentList[["NatAbuBase"]] <- unlist(NatAbuBaseList)
        tmpFragmentList[["MassShiftTracer"]] <- unlist(MassShiftTracerList)
      } else {
        if(verbose){message(date(), " :: NO TRACER FOR MOLECULE ", MoleculeInformation[1,1], " AND FRAGMENT #", Fragment)}
      }  #NumberTracers>0
    }  #if(UltraHighRes==1)
    
    FragmentList[[Fragment]] <- tmpFragmentList
  }  # Fragment
  
  names(FragmentList) <- stringr::str_c("Fragment_", rep(seq_len(NumberFragments)))
  
  return(FragmentList)
  
}