a b/data/data_reading_SEERBOMB.r 

  1
  
setwd(".")





  
  

  2
  
options(stringsAsFactors = FALSE)





  
  

  3
  





  
  

  4
  
# list.of.packages <- c("PRROC", "e1071", "randomForest","class", "gmodels", "formula.tools")





  
  

  5
  





  
  

  6
  
list.of.packages <- c("PRROC", "e1071", "randomForest","class", "gmodels", "SEERaBomb", "miceadds")





  
  

  7
  
new.packages <- list.of.packages[!(list.of.packages %in% installed.packages()[,"Package"])]





  
  

  8
  
if(length(new.packages)) install.packages(new.packages)





  
  

  9
  





  
  

  10
  





  
  

  11
  
# library("PRROC")





  
  

  12
  
# library("e1071")





  
  

  13
  
# library("randomForest")





  
  

  14
  
# library("class")





  
  

  15
  
# library("gmodels")





  
  

  16
  
library("SEERaBomb")





  
  

  17
  
library("miceadds")





  
  

  18
  





  
  

  19
  
num_to_return <- 1





  
  

  20
  
exe_num <- sample(1:as.numeric(Sys.time()), num_to_return)





  
  

  21
  





  
  

  22
  
# https://rdrr.io/cran/SEERaBomb/src/inst/doc/examples/mkDataBinaries.R





  
  

  23
  





  
  

  24
  
df <- getFields(seerHome="./SEER_1973_2015_CUSTOM_TEXTDATA/")





  
  

  25
  





  
  

  26
  
rdf=pickFields(df)





  
  

  27
  
# pickFields(df, picks=c("casenum","reg","race","sex","agedx",  "yrbrth","seqnum","modx","yrdx","histo3",      "ICD9","COD","surv","radiatn","chemo", "tvalue"))





  
  

  28
  





  
  

  29
  
## Dataset with date of birth





  
  

  30
  





  
  

  31
  
rdf_TumorSize_Age=pickFields(df, picks=c("casenum","reg","race","sex","agedx",  "yrbrth","seqnum","modx","yrdx","histo3",   "ICD9","COD","surv","radiatn","chemo", "tvalue", "mvalue"))





  
  

  32
  





  
  

  33
  
mkSEER(rdf_TumorSize_Age, seerHome="./SEER_1973_2015_CUSTOM_TEXTDATA/")





  
  

  34
  





  
  

  35
  
load.Rdata("./SEER_1973_2015_CUSTOM_TEXTDATA/mrgd/cancDef_ALL_TumorSize_Metastasis.RData", "data_ALL_TMY")





  
  

  36
  





  
  

  37
  
lung_cancer_data <- data_ALL_TMY%>%filter(cancer=="lung")





  
  

  38
  
lung_cancer_dataframe <- as.data.frame(lung_cancer_data)





  
  

  39
  
colnames(lung_cancer_dataframe)





  
  

  40
  





  
  

  41
  
lung_cancer_dataframe_TMY <- lung_cancer_dataframe[,c("tvalue","yrbrth","mvalue")]





  
  

  42
  





  
  

  43
  
# Removes the NA values





  
  

  44
  
lung_cancer_dataframe_TMY_complete_cases <- lung_cancer_dataframe_TMY[complete.cases(lung_cancer_dataframe_TMY), ]





  
  

  45
  





  
  

  46
  





  
  

  47
  
# AJCC 3rd Edition, TNM, and Stage in SEER Data





  
  

  48
  
# https://seer.cancer.gov/seerstat/variables/seer/ajcc-stage/3rd.html





  
  

  49
  





  
  

  50
  
# Let's keep only M0 (code 00) and M1 (code 10) 





  
  

  51
  
lung_cancer_dataframe_TMY_noNA_onlyM0M1 <- lung_cancer_dataframe_TMY_complete_cases[lung_cancer_dataframe_TMY_complete_cases$"mvalue" == "10" | lung_cancer_dataframe_TMY_complete_cases$"mvalue" == "0", ]





  
  

  52
  





  
  

  53
  
colnames(lung_cancer_dataframe_TMY_noNA_onlyM0M1) <- c("TumorSize", "YearOfBirth", "Metastasis")





  
  

  54
  





  
  

  55
  
 write.table(lung_cancer_dataframe_TMY_noNA_onlyM0M1, paste("lung_cancer_dataframe_TMY_noNA_onlyM0M1_time", exe_num, ".csv", sep=""),col.names=TRUE, row.names=FALSE, sep=",")





  
  

  56
  
 





  
  

  57
  
## Dataset with "agerec"





  
  

  58
  





  
  

  59
  
rdf_TumorSize_AgeB=pickFields(df, picks=c("casenum","reg","race","sex","agedx",  "agerec","seqnum","modx","yrdx","histo3",   "ICD9","COD","surv","radiatn","chemo", "tvalue", "mvalue"))





  
  

  60
  





  
  

  61
  
mkSEER(rdf_TumorSize_AgeB, seerHome="./SEER_1973_2015_CUSTOM_TEXTDATA/", outFile="cancDef_ALL_TumorSize_Age_Metastasis")





  
  

  62
  





  
  

  63
  
load.Rdata("./SEER_1973_2015_CUSTOM_TEXTDATA/mrgd/cancDef_ALL_TumorSize_Age_Metastasis.RData", "data_ALL_TMA")





  
  

  64
  





  
  

  65
  
lung_cancer_age_data <- data_ALL_TMA%>%filter(cancer=="lung")





  
  

  66
  
lung_cancer_age_dataframe <- as.data.frame(lung_cancer_age_data)





  
  

  67
  
colnames(lung_cancer_age_dataframe)





  
  

  68
  





  
  

  69
  
lung_cancer_dataframe_TMA <- lung_cancer_age_dataframe[,c("tvalue","agedx","mvalue")]





  
  

  70
  





  
  

  71
  
# Removes the NA values





  
  

  72
  
lung_cancer_dataframe_TMA_complete_cases <- lung_cancer_dataframe_TMA[complete.cases(lung_cancer_dataframe_TMA), ]





  
  

  73
  





  
  

  74
  





  
  

  75
  
# AJCC 3rd Edition, TNM, and Stage in SEER Data





  
  

  76
  
# https://seer.cancer.gov/seerstat/variables/seer/ajcc-stage/3rd.html





  
  

  77
  





  
  

  78
  
# Let's keep only M0 (code 00) and M1 (code 10) 





  
  

  79
  
lung_cancer_dataframe_TMA_noNA_onlyM0M1 <- lung_cancer_dataframe_TMA_complete_cases[lung_cancer_dataframe_TMA_complete_cases$"mvalue" == "10" | lung_cancer_dataframe_TMA_complete_cases$"mvalue" == "0", ]





  
  

  80
  





  
  

  81
  
colnames(lung_cancer_dataframe_TMA_noNA_onlyM0M1) <- c("TumorSize", "AgeAtDiagnosis", "Metastasis")





  
  

  82
  
lung_cancer_dataframe_TMA_noNA_onlyM0M1$Metastasis <- lung_cancer_dataframe_TMA_noNA_onlyM0M1$Metastasis/10





  
  

  83
  





  
  

  84
  
 write.table(lung_cancer_dataframe_TMA_noNA_onlyM0M1, paste("lung_cancer_dataframe_TMA_noNA_onlyM0M1_time", exe_num, ".csv", sep=""),col.names=TRUE, row.names=FALSE, sep=",")





  
  

  85
  
 





  
  

  86