# Wild type sequence provided in the "Dataset Description":
wtseq <- 'VPVNPEPDATSVENVALKTGSGDSQSDPIKADLEVKGQSALPFDVDCWAILCKGAPNVLQRVNEKTKNSNRDRSGANKGPFKDPQKWGIKALPPKNPSWSAQDFKSPEEYAFASSLQGGTNAILAPVNLASQNSQGGVLNGFYSANKVAQFDPSKPQQTKGTWFQITKFTGAAGPYCKALGSNDKSVCDKNKNIAGDWGFDPAKWAYQYDEKNNKFNYVGK'

# Read testing set sequences and pH:
test <- read.csv('../input/novozymes-enzyme-stability-prediction/test.csv')

# Add mutation information to testing set:
test[,c('type','resid','wt','mut')] <- do.call(rbind,lapply(test$protein_sequence,function(seq){
  # case 1 = wild type:
  if(seq==wtseq){ 
    return(c('WT',-1,'_','_'))
  # case 2 = substitution:
  } else if(nchar(seq)==nchar(wtseq)){ 
    i <- mapply(function(x,y) which(x!=y)[1], strsplit(seq,""), strsplit(wtseq,""))
    return(c('SUB',i,substr(wtseq,i,i),substr(seq,i,i)))
  # case 3 = deletion:
  } else if(nchar(seq)<nchar(wtseq)){ 
    wtsub <- substr(wtseq,1,nchar(seq))
    i <- mapply(function(x,y) which(x!=y)[1], strsplit(seq,""), strsplit(wtsub,""))
    return(c('DEL',i,substr(wtseq,i,i),'_'))
  }
}))
head(test)