NejiVocabularies

Vocabularies

The vocabulary used in the system was created based on the concepts available in the UMLS, using the following files:

• MRCONSO.RRF for extracting the concepts
• MRSTY.RRF for extracting the UMLS semantic type associated with each concept

To obtain access to the MRCONSO.RRF and MRSTY.RRF resource files, please request your access here.

To create the necessary vocabulary files from the UMLS, the following command is used (in Linux systems):

cat MRCONSO.RRF | grep "RXNORM" > RXNORM.csv
cat MRCONSO.RRF | grep "DRUGBANK" > DRUGBANK.csv
cat MRCONSO.RRF | grep "|AOD|" > AOD_unfiltered.csv

Despite the specificity of RxNorm and DrugBank dictionaries, the AOD dictionary covers a broader spectrum of concepts. To filter the AOD_unfiltered.csv vocabulary file so that it only contains UMLS concepts belonging to the "Chemicals & Drugs" semantic group, the following command must be run (in Linux systems), where all Txxx codes correspond to semantic types within the "Chemicals & Drugs" semantic group:

cat AOD_unfiltered.csv | grep -E "T116|T195|T123|T122|T103|T120|T104|T200|T196|T126|T131|T125|T129|T130|T197|T114|T109|T121|T192|T127" > AOD.csv

NOTE: if the resulting vocabulary files are not stored in vocabularies/, it is necessary to configure the [vocabularies] variables in Settings.ini with their corresponding paths