README.Md

In dlindholm/oboe: Open Biomedical Ontology Extraction Tools (oboe)

oboe (Open Biomedical Ontologies Extraction tools) package

Daniel Lindholm

9/14/2016

This package reads ontologies in OBO format and parses terms into a list in R. It could be used to find where in the tree a specific term is located, and has functionality to map cross-references (such as ICD9, ICD10, OMIM, SNOMED, whatever cross-reference is in the database) to an ontology identifier.

Below is a small demo on how it could be used. For this example, Disease Ontology was used, but it should work with any ontology saved in the OBO format (haven’t tested yet, though).

Initial steps

First, load package into R:

if(!"devtools" %in% installed.packages()) install.packages("devtools")
if(!"oboe" %in% installed.packages()) devtools::install_github("dlindholm/oboe")
library("oboe")

## Loading required package: parallel

Then read OBO file, in this case from an URL to get the latest disease ontology database, but this could of course also be a local file:

doid <- read_obo("http://purl.obolibrary.org/obo/doid.obo")

This parses the OBO file into a list:

head(doid)

## [[1]]
## [[1]]$id
## [1] "DOID:0001816"
## 
## [[1]]$name
## [1] "angiosarcoma"
## 
## [[1]]$alt_id
## [1] "DOID:267"  "DOID:4508"
## 
## [[1]]$def
## [1] "A malignant Vascular tumor that results_in rapidly proliferating, extensively infiltrating anaplastic cells derived_from blood vessels and derived_from the lining of irregular blood-filled spaces. [url:http\\://emedicine.medscape.com/article/276512-overview, url:http\\://en.wikipedia.org/wiki/Hemangiosarcoma, url:http\\://www.ncbi.nlm.nih.gov/pubmed/23327728]"
## 
## [[1]]$subset
## [1] "NCIthesaurus"
## 
## [[1]]$synonym
## [1] "hemangiosarcoma EXACT []"
## 
## [[1]]$xref
## [1] "MSH:D006394"                      "NCI:C3088"                       
## [3] "NCI:C9275"                        "SNOMEDCT_US_2016_03_01:33176006" 
## [5] "SNOMEDCT_US_2016_03_01:39000009"  "SNOMEDCT_US_2016_03_01:403977003"
## [7] "UMLS_CUI:C0018923"                "UMLS_CUI:C0854893"               
## 
## [[1]]$is_a
## [1] "DOID:1115 ! sarcoma"
## 
## 
## [[2]]
## [[2]]$id
## [1] "DOID:0002116"
## 
## [[2]]$name
## [1] "pterygium"
## 
## [[2]]$synonym
## [1] "surfer's eye EXACT []"
## 
## [[2]]$xref
##  [1] "ICD10CM:H11.0"                    "ICD10CM:H11.00"                  
##  [3] "ICD10CM:H11.009"                  "ICD9CM:372.4"                    
##  [5] "ICD9CM:372.40"                    "MSH:D011625"                     
##  [7] "SNOMEDCT_US_2016_03_01:155165000" "SNOMEDCT_US_2016_03_01:193879003"
##  [9] "SNOMEDCT_US_2016_03_01:193884009" "SNOMEDCT_US_2016_03_01:77489003" 
## [11] "UMLS_CUI:C0033999"               
## 
## [[2]]$is_a
## [1] "DOID:10124 ! corneal disease"
## 
## [[2]]$created_by
## [1] "laronhughes"
## 
## [[2]]$creation_date
## [1] "2010-06-30T02:44:30Z"
## 
## 
## [[3]]
## [[3]]$id
## [1] "DOID:0014667"
## 
## [[3]]$name
## [1] "disease of metabolism"
## 
## [[3]]$def
## [1] "A disease that involving errors in metabolic processes of building or degradation of molecules. [url:http\\://www.ncbi.nlm.nih.gov/books/NBK22259/]"
## 
## [[3]]$subset
## [1] "NCIthesaurus"
## 
## [[3]]$synonym
## [1] "metabolic disease  EXACT [SNOMEDCT_2005_07_31:75934005]"
## 
## [[3]]$xref
##  [1] "ICD10CM:E88.9"                    "ICD9CM:277.9"                    
##  [3] "MSH:D008659"                      "NCI:C3235"                       
##  [5] "SNOMEDCT_US_2016_03_01:154733004" "SNOMEDCT_US_2016_03_01:190961002"
##  [7] "SNOMEDCT_US_2016_03_01:267456000" "SNOMEDCT_US_2016_03_01:30390004" 
##  [9] "SNOMEDCT_US_2016_03_01:75934005"  "UMLS_CUI:C0025517"               
## 
## [[3]]$is_a
## [1] "DOID:4 ! disease"
## 
## 
## [[4]]
## [[4]]$id
## [1] "DOID:0050001"
## 
## [[4]]$name
## [1] "Actinomadura madurae infectious disease"
## 
## [[4]]$subset
## [1] "gram-positive_bacterial_infectious_disease"
## 
## [[4]]$is_obsolete
## [1] "true"
## 
## 
## [[5]]
## [[5]]$id
## [1] "DOID:0050002"
## 
## [[5]]$name
## [1] "Actinomadura pelletieri infectious disease"
## 
## [[5]]$subset
## [1] "gram-positive_bacterial_infectious_disease"
## 
## [[5]]$is_obsolete
## [1] "true"
## 
## 
## [[6]]
## [[6]]$id
## [1] "DOID:0050003"
## 
## [[6]]$name
## [1] "Streptomyces somaliensis infectious disease"
## 
## [[6]]$subset
## [1] "gram-positive_bacterial_infectious_disease"
## 
## [[6]]$is_obsolete
## [1] "true"

You could then remove the obsolete terms:

doid <- remove_obsolete(doid)

### Finding terms You could look up terms in different ways. The *find\_entry* function is used for that. Look for a specific entry id like this: wzxhzdk:4 ## [[1]] ## [[1]]$id ## [1] "DOID:8805" ## ## [[1]]$name ## [1] "intermediate coronary syndrome" ## ## [[1]]$synonym ## [1] "(Angina: [crescendo] or [unstable] or [at rest]) or (preinfarction syndrome) or (impending infarction) EXACT [SNOMEDCT_2005_07_31:194814006]" ## [2] "Angina at rest EXACT [SNOMEDCT_2005_07_31:233818002]" ## [3] "Angina at rest EXACT [SNOMEDCT_2005_07_31:194817004]" ## [4] "Angina at rest EXACT [SNOMEDCT_2005_07_31:155313008]" ## [5] "Anginal chest pain at rest EXACT [SNOMEDCT_2005_07_31:59021001]" ## [6] "Impending infarction (disorder) EXACT [SNOMEDCT_2005_07_31:25106000]" ## [7] "Preinfarction angina EXACT [MTHICD9_2006:411.1]" ## [8] "Preinfarction angina (disorder) EXACT [SNOMEDCT_2005_07_31:64333001]" ## [9] "Unstable angina EXACT [SNOMEDCT_2005_07_31:155308009]" ## [10] "Unstable angina EXACT [MTH:NOCODE]" ## [11] "Unstable angina EXACT [SNOMEDCT_2005_07_31:194816008]" ## [12] "Unstable angina EXACT [SNOMEDCT_2005_07_31:233820004]" ## [13] "Worsening angina EXACT [SNOMEDCT_2005_07_31:4557003]" ## [14] "Worsening angina EXACT [SNOMEDCT_2005_07_31:155314002]" ## [15] "Worsening angina EXACT [SNOMEDCT_2005_07_31:194819001]" ## ## [[1]]$xref ## [1] "ICD10CM:I20.0" "ICD9CM:411.1" ## [3] "MSH:D000789" "NCI:C66911" ## [5] "SNOMEDCT_US_2016_03_01:155308009" "SNOMEDCT_US_2016_03_01:155313008" ## [7] "SNOMEDCT_US_2016_03_01:155314002" "SNOMEDCT_US_2016_03_01:194814006" ## [9] "SNOMEDCT_US_2016_03_01:194816008" "SNOMEDCT_US_2016_03_01:194817004" ## [11] "SNOMEDCT_US_2016_03_01:194819001" "SNOMEDCT_US_2016_03_01:233818002" ## [13] "SNOMEDCT_US_2016_03_01:233820004" "SNOMEDCT_US_2016_03_01:25106000" ## [15] "SNOMEDCT_US_2016_03_01:4557003" "SNOMEDCT_US_2016_03_01:59021001" ## [17] "SNOMEDCT_US_2016_03_01:64333001" "UMLS_CUI:C0002965" ## ## [[1]]$is_a ## [1] "DOID:3393 ! coronary artery disease" or for a specific name: wzxhzdk:5 ## [[1]] ## [[1]]$id ## [1] "DOID:10763" ## ## [[1]]$name ## [1] "hypertension" ## ## [[1]]$def ## [1] "An artery disease characterized by chronic elevated blood pressure in the arteries. [PMID:24352797, url:https\\://en.wikipedia.org/wiki/Hypertension]" ## ## [[1]]$synonym ## [1] "High blood pressure (& [essential hypertension]) EXACT [SNOMEDCT_2005_07_31:194757006]" ## [2] "HTN EXACT [CSP2005:0571-5243]" ## [3] "hyperpiesia EXACT [CSP2005:4003-0017]" ## [4] "hypertensive disease RELATED [MTH:NOCODE]" ## [5] "vascular hypertensive disorder EXACT [NCI2004_11_17:C3117]" ## ## [[1]]$xref ## [1] "EFO:0000537" "ICD10CM:I10" ## [3] "ICD10CM:I10-I15" "ICD9CM:401-405.99" ## [5] "ICD9CM:997.91" "MSH:D006973" ## [7] "NCI:C3117" "SNOMEDCT_US_2016_03_01:155295004" ## [9] "SNOMEDCT_US_2016_03_01:155302005" "SNOMEDCT_US_2016_03_01:194756002" ## [11] "SNOMEDCT_US_2016_03_01:194757006" "SNOMEDCT_US_2016_03_01:194760004" ## [13] "SNOMEDCT_US_2016_03_01:194794002" "SNOMEDCT_US_2016_03_01:195537001" ## [15] "SNOMEDCT_US_2016_03_01:266287006" "SNOMEDCT_US_2016_03_01:38341003" ## [17] "UMLS_CUI:C0020538" ## ## [[1]]$is_a ## [1] "DOID:0050828 ! artery disease" ### Mapping other codes to an ontology ID If you’d like to map a specific code from another classification, it is possible to do so if this classification is in the *xref* part of the database. You could find corresponding ontology ID’s for ICD10 (which is the function’s default): wzxhzdk:6 ## [1] "DOID:6000" or from other classifications: wzxhzdk:7 ## [1] "DOID:10763" wzxhzdk:8 ## [1] "DOID:12930" wzxhzdk:9 ## [1] "DOID:9408" wzxhzdk:10 ### Finding a code’s position in the tree Too look up where in the hierarchy an entry is, you could do like this: wzxhzdk:11 ## id text depth ## 1 DOID:4 disease 1 ## 2 DOID:7 disease of anatomical entity 2 ## 3 DOID:1287 cardiovascular system disease 3 ## 4 DOID:178 vascular disease 4 ## 5 DOID:0050828 artery disease 5 ## 6 DOID:3393 coronary artery disease 6 ## 7 DOID:5844 myocardial infarction 7 ## 8 DOID:9408 acute myocardial infarction 8 Or, if you only want the position, use *get\_depth*: wzxhzdk:12 ## [1] 8 ### Next steps Using these functions, you could convert an ICD10 code into an ontology ID, and further abstract the diagnosis to the desired level of detail, and thus group diagnoses. This is an early version of the package, which will be developed further to add additional functionality, and to test with other ontologies in the OBO format. Any input and suggestions, or collaboration via GitHub is greatly appreciated!

dlindholm/oboe documentation built on May 15, 2019, 9:18 a.m.