Processing ehrlink

library(dplyr)
library(DT)
library(ggplot2)

options(stringsAsFactors=FALSE)

# read ehrlink problem-medication pairs
ehrlink.df <- file.path('download', 'amiajnl-2012-000852-s1.tsv') %>%
  read.delim()

# read rxnorm concept to rxnorm active ingredient mapping
ingredient_map.df <- file.path('data', 'rxnorm-as-ingredient.tsv') %>%
  read.delim()

# read ehrlink medication to rxnorm mapping
medication_map.df <- 'https://raw.githubusercontent.com/antoine-lizee/RRxNorm/531bfcb892ca82ba3a60e9ba42b09373977f08ef/Output/allResolvedMatches.csv' %>%
  read.csv(na.strings = '')

# read ehrlink problem to DO (disease ontology) mapping
problem_map.df <- file.path('data', 'problem-to-doid.tsv') %>%
  read.delim(na.strings = '')

mapped.df <- ehrlink.df %>% dplyr::inner_join(
  medication_map.df %>%
    dplyr::filter(score >= 55) %>%
    dplyr::select(medication_definition_id, rxcui)
) %>% dplyr::inner_join(
  ingredient_map.df %>%
    dplyr::filter(n_ingredients == 1) %>%
    dplyr::transmute(rxcui = input_rxcui, ingredient_rxcui, ingredient_name)
) %>% dplyr::inner_join(
  problem_map.df %>%
    dplyr::filter(! is.na(doid)) %>%
    dplyr::distinct(problem_definition_id, doid) %>%
    dplyr::transmute(problem_definition_id, doid_code = doid, doid_name = name)
)

## Joining by: "medication_definition_id"
## Joining by: "rxcui"
## Joining by: "problem_definition_id"

mapped.df %>%
  write.table(file.path('data', 'indications.tsv'), quote=FALSE, row.names=FALSE, sep='\t')

mapped.df %>% DT::datatable()

# DT::datatable(ehrlink.df)
# DT::datatable(ingredient_map.df)
# DT::datatable(medication_map.df)
# DT::datatable(problem_map.df)
# DT::datatable(doslim_map.df)