Using UMLS Concepts with MeSH

The Medical Subject Headings (MeSH) terms returned from a PubMed search can be further analyzed by mapping them to Unified Medical Language System (UMLS) concepts, as well as filtering the MeSH Terms by concepts.

For both mapping MeSH to UMLS Concepts and filtering MeSH by concept, the following backends are supported:

MySQL
SQLite
DataFrames

Set Up

using SQLite
using MySQL
using BioMedQuery.DBUtils
using BioMedQuery.Processes
using BioServices.UMLS
using BioMedQuery.PubMed
using DataFrames

Credentials are environment variables (e.g set in your .juliarc.jl)

umls_user = ENV["UMLS_USER"];
umls_pswd = ENV["UMLS_PSSWD"];
email = ""; # Only needed if you want to contact NCBI with inqueries
search_term = """(obesity[MeSH Major Topic]) AND ("2010"[Date - Publication] : "2012"[Date - Publication])""";
umls_concept = "Disease or Syndrome";
max_articles = 5;
results_dir = ".";
verbose = true;

results_dir = ".";

"."

MySQL

Map Medical Subject Headings (MeSH) to UMLS

This example demonstrates the typical workflow to populate a MESH2UMLS database table relating all concepts associated with all MeSH terms in the input database.

Note: this example reuses the MySQL DB from the PubMed Search and Save example.

Create MySQL DB connection

host = "127.0.0.1";
mysql_usr = "root";
mysql_pswd = "";
dbname = "pubmed_obesity_2010_2012";


db_mysql = MySQL.connect(host, mysql_usr, mysql_pswd, db = dbname);

MySQL Connection
------------
Host: 127.0.0.1
Port: 3306
User: root
DB:   pubmed_obesity_2010_2012

Map MeSH to UMLS

@time map_mesh_to_umls_async!(db_mysql, umls_user, umls_pswd; append_results=false, timeout=3);

┌ Warning: `getindex(df::DataFrame, col_ind::ColumnIndex)` is deprecated, use `df[!, col_ind]` instead.
│   caller = #map_mesh_to_umls_async!#7(::Int64, ::Bool, ::Bool, ::Int64, ::Function, ::MySQL.Connection, ::String, ::String) at pubmed_mesh_to_umls_map.jl:44
└ @ BioMedQuery.Processes ~/build/bcbi/BioMedQuery.jl/src/Processes/pubmed_mesh_to_umls_map.jl:44
----------Matching MESH to UMLS-----------
["Adult", "Aged", "Aged, 80 and over", "Analysis of Variance", "Body Weight", "C-Reactive Protein", "Child", "Cross-Sectional Studies", "Fatigue", "Female", "Fibromyalgia", "Germany", "Health Status", "Humans", "Japan", "Male", "Middle Aged", "Nutrition Surveys", "Obesity", "Pain", "Pain Measurement", "Physical Fitness", "Prognosis", "Quality of Life", "Surveys and Questionnaires", "Reference Values", "Risk Factors", "ROC Curve", "Severity of Illness Index", "Sports", "Television", "Thyrotropin", "Biomarkers", "Weight Gain", "Exercise", "Body Mass Index", "Incidence", "Prevalence", "Logistic Models", "Odds Ratio", "Case-Control Studies", "Age Distribution", "Sex Distribution", "Sleep Apnea, Obstructive", "Metabolic Syndrome", "Overweight", "Waist Circumference", "Young Adult", "Obesity, Abdominal", "Republic of Korea", "Sedentary Behavior", "Pediatric Obesity"]
[ Info: UTS: Requesting new TGT
[ Info: Descriptor 1 out of 52: Adult
[ Info: Descriptor 31 out of 52: Television
[ Info: Descriptor 36 out of 52: Body Mass Index
[ Info: Descriptor 35 out of 52: Exercise
[ Info: Descriptor 32 out of 52: Thyrotropin
[ Info: Descriptor 30 out of 52: Sports
[ Info: Descriptor 29 out of 52: Severity of Illness Index
[ Info: Descriptor 34 out of 52: Weight Gain
[ Info: Descriptor 40 out of 52: Odds Ratio
[ Info: Descriptor 44 out of 52: Sleep Apnea, Obstructive
[ Info: Descriptor 39 out of 52: Logistic Models
[ Info: Descriptor 43 out of 52: Sex Distribution
[ Info: Descriptor 42 out of 52: Age Distribution
[ Info: Descriptor 37 out of 52: Incidence
[ Info: Descriptor 33 out of 52: Biomarkers
[ Info: Descriptor 41 out of 52: Case-Control Studies
[ Info: Descriptor 9 out of 52: Fatigue
[ Info: Descriptor 4 out of 52: Analysis of Variance
[ Info: Descriptor 3 out of 52: Aged, 80 and over
[ Info: Descriptor 6 out of 52: C-Reactive Protein
[ Info: Descriptor 7 out of 52: Child
[ Info: Descriptor 5 out of 52: Body Weight
[ Info: Descriptor 46 out of 52: Overweight
[ Info: Descriptor 2 out of 52: Aged
[ Info: Descriptor 15 out of 52: Japan
[ Info: Descriptor 10 out of 52: Female
[ Info: Descriptor 12 out of 52: Germany
[ Info: Descriptor 14 out of 52: Humans
[ Info: Descriptor 8 out of 52: Cross-Sectional Studies
[ Info: Descriptor 21 out of 52: Pain Measurement
[ Info: Descriptor 11 out of 52: Fibromyalgia
[ Info: Descriptor 13 out of 52: Health Status
[ Info: Descriptor 18 out of 52: Nutrition Surveys
[ Info: Descriptor 17 out of 52: Middle Aged
[ Info: Descriptor 20 out of 52: Pain
[ Info: Descriptor 19 out of 52: Obesity
[ Info: Descriptor 28 out of 52: ROC Curve
[ Info: Descriptor 23 out of 52: Prognosis
[ Info: Descriptor 25 out of 52: Surveys and Questionnaires
[ Info: Descriptor 16 out of 52: Male
[ Info: Descriptor 26 out of 52: Reference Values
[ Info: Descriptor 22 out of 52: Physical Fitness
[ Info: Descriptor 27 out of 52: Risk Factors
[ Info: Descriptor 38 out of 52: Prevalence
[ Info: Descriptor 24 out of 52: Quality of Life
[ Info: Descriptor 50 out of 52: Republic of Korea
[ Info: Descriptor 49 out of 52: Obesity, Abdominal
[ Info: Descriptor 51 out of 52: Sedentary Behavior
[ Info: Descriptor 47 out of 52: Waist Circumference
[ Info: Descriptor 45 out of 52: Metabolic Syndrome
[ Info: Descriptor 48 out of 52: Young Adult
[ Info: Descriptor 52 out of 52: Pediatric Obesity
[ Info: Descriptor 51 out of 52: Sedentary Behavior
 11.464593 seconds (8.18 M allocations: 407.230 MiB, 2.39% gc time)

Explore the output table

db_query(db_mysql, "SELECT * FROM mesh2umls")

56 rows × 2 columns

	mesh	umls
	String	String
1	Adult	Age Group
2	Age Distribution	Quantitative Concept
3	Aged	Organism Attribute
4	Aged, 80 and over	Age Group
5	Analysis of Variance	Quantitative Concept
6	Biomarkers	Clinical Attribute
7	Body Mass Index	Diagnostic Procedure
8	Body Weight	Organism Attribute
9	C-Reactive Protein	Amino Acid, Peptide, or Protein
10	C-Reactive Protein	Immunologic Factor
11	Case-Control Studies	Research Activity
12	Child	Age Group
13	Cross-Sectional Studies	Research Activity
14	Exercise	Daily or Recreational Activity
15	Fatigue	Sign or Symptom
16	Female	Population Group
17	Fibromyalgia	Disease or Syndrome
18	Germany	Geographic Area
19	Health Status	Qualitative Concept
20	Humans	Human
21	Incidence	Quantitative Concept
22	Japan	Geographic Area
23	Logistic Models	Intellectual Product
24	Logistic Models	Quantitative Concept
25	Male	Organism Attribute
26	Metabolic Syndrome	Disease or Syndrome
27	Middle Aged	Age Group
28	Nutrition Surveys	Research Activity
29	Obesity	Disease or Syndrome
30	Obesity, Abdominal	Finding
31	Odds Ratio	Quantitative Concept
32	Overweight	Finding
33	Pain	Food
34	Pain Measurement	Diagnostic Procedure
35	Pediatric Obesity	Disease or Syndrome
36	Physical Fitness	Idea or Concept
37	Prevalence	Quantitative Concept
38	Prognosis	Health Care Activity
39	Quality of Life	Idea or Concept
40	Reference Values	Quantitative Concept
41	Republic of Korea	Geographic Area
42	Risk Factors	Finding
43	ROC Curve	Quantitative Concept
44	Sedentary Behavior	Finding
45	Severity of Illness Index	Quantitative Concept
46	Sex Distribution	Quantitative Concept
47	Sleep Apnea, Obstructive	Disease or Syndrome
48	Sports	Daily or Recreational Activity
49	Surveys and Questionnaires	Research Activity
50	Television	Manufactured Object
51	Thyrotropin	Amino Acid, Peptide, or Protein
52	Thyrotropin	Hormone
53	Thyrotropin	Pharmacologic Substance
54	Waist Circumference	Clinical Attribute
55	Weight Gain	Finding
56	Young Adult	Age Group

Filtering MeSH terms by UMLS concept

Getting the descriptor to index dictionary and the occurence matrix

@time labels2ind, occur = umls_semantic_occurrences(db_mysql, umls_concept);

(Dict("Obesity"=>1,"Pediatric Obesity"=>2,"Sleep Apnea, Obstructive"=>3,"Metabolic Syndrome"=>4,"Fibromyalgia"=>5), 
  [1, 1]  =  1.0
  [5, 1]  =  1.0
  [1, 2]  =  1.0
  [3, 2]  =  1.0
  [2, 3]  =  1.0
  [1, 4]  =  1.0
  [4, 5]  =  1.0)

Descriptor to Index Dictionary

labels2ind

Dict{String,Int64} with 5 entries:
  "Obesity"                  => 1
  "Pediatric Obesity"        => 2
  "Sleep Apnea, Obstructive" => 3
  "Metabolic Syndrome"       => 4
  "Fibromyalgia"             => 5

Output Data Matrix

Matrix(occur)

5×5 Array{Float64,2}:
 1.0  1.0  0.0  1.0  0.0
 0.0  0.0  1.0  0.0  0.0
 0.0  1.0  0.0  0.0  0.0
 0.0  0.0  0.0  0.0  1.0
 1.0  0.0  0.0  0.0  0.0

SQLite

This example demonstrates the typical workflow to populate a MESH2UMLS database table relating all concepts associated with all MeSH terms in the input database.

Note: this example reuses the SQLite DB from the PubMed Search and Save example.

Create SQLite DB connection

db_path = "$(results_dir)/pubmed_obesity_2010_2012.db";
db_sqlite = SQLite.DB(db_path);

Getting 5 articles, starting at index 0
------ESearch--------
------EFetch--------
------Save to database--------
Saving 5 articles to database
┌ Warning: `getindex(df::DataFrame, col_ind::ColumnIndex)` is deprecated, use `df[!, col_ind]` instead.
│   caller = select_columns at sqlite_db_utils.jl:9 [inlined]
└ @ Core ~/build/bcbi/BioMedQuery.jl/src/DBUtils/sqlite_db_utils.jl:9
┌ Warning: `getindex(df::DataFrame, col_ind::ColumnIndex)` is deprecated, use `df[!, col_ind]` instead.
│   caller = insert_row!(::SQLite.DB, ::String, ::Dict{Symbol,Any}, ::Bool) at sqlite_db_utils.jl:59
└ @ BioMedQuery.DBUtils ~/build/bcbi/BioMedQuery.jl/src/DBUtils/sqlite_db_utils.jl:59
Finished searching, total number of articles: 5

Map MeSH to UMLS

@time map_mesh_to_umls_async!(db_sqlite, umls_user, umls_pswd; append_results=false, timeout=3);

┌ Warning: `getindex(df::DataFrame, col_ind::ColumnIndex)` is deprecated, use `df[!, col_ind]` instead.
│   caller = #map_mesh_to_umls_async!#7(::Int64, ::Bool, ::Bool, ::Int64, ::Function, ::SQLite.DB, ::String, ::String) at pubmed_mesh_to_umls_map.jl:44
└ @ BioMedQuery.Processes ~/build/bcbi/BioMedQuery.jl/src/Processes/pubmed_mesh_to_umls_map.jl:44
----------Matching MESH to UMLS-----------
Union{Missing, String}["Reference Values", "Republic of Korea", "ROC Curve", "Fatigue", "Obesity", "Risk Factors", "Logistic Models", "Severity of Illness Index", "Male", "Case-Control Studies", "Analysis of Variance", "Sedentary Behavior", "Prevalence", "Quality of Life", "Odds Ratio", "Exercise", "Body Mass Index", "Aged", "Child", "Sex Distribution", "Adult", "Germany", "Sports", "Thyrotropin", "Pediatric Obesity", "Humans", "Japan", "Cross-Sectional Studies", "Weight Gain", "Middle Aged", "Surveys and Questionnaires", "Health Status", "Young Adult", "Incidence", "Prognosis", "Body Weight", "Pain Measurement", "Waist Circumference", "Metabolic Syndrome", "Pain", "Nutrition Surveys", "Fibromyalgia", "Sleep Apnea, Obstructive", "Television", "Age Distribution", "Overweight", "Physical Fitness", "Female", "Biomarkers", "Obesity, Abdominal", "C-Reactive Protein", "Aged, 80 and over"]
[ Info: UTS: Reading TGT from file
[ Info: Descriptor 5 out of 52: Obesity
[ Info: Descriptor 4 out of 52: Fatigue
[ Info: Descriptor 8 out of 52: Severity of Illness Index
[ Info: Descriptor 1 out of 52: Reference Values
[ Info: Descriptor 9 out of 52: Male
[ Info: Descriptor 6 out of 52: Risk Factors
[ Info: Descriptor 12 out of 52: Sedentary Behavior
[ Info: Descriptor 10 out of 52: Case-Control Studies
[ Info: Descriptor 16 out of 52: Exercise
[ Info: Descriptor 11 out of 52: Analysis of Variance
[ Info: Descriptor 13 out of 52: Prevalence
[ Info: Descriptor 14 out of 52: Quality of Life
[ Info: Descriptor 15 out of 52: Odds Ratio
[ Info: Descriptor 24 out of 52: Thyrotropin
[ Info: Descriptor 23 out of 52: Sports
[ Info: Descriptor 22 out of 52: Germany
[ Info: Descriptor 26 out of 52: Humans
[ Info: Descriptor 20 out of 52: Sex Distribution
[ Info: Descriptor 17 out of 52: Body Mass Index
[ Info: Descriptor 21 out of 52: Adult
[ Info: Descriptor 19 out of 52: Child
[ Info: Descriptor 25 out of 52: Pediatric Obesity
[ Info: Descriptor 18 out of 52: Aged
[ Info: Descriptor 27 out of 52: Japan
[ Info: Descriptor 28 out of 52: Cross-Sectional Studies
[ Info: Descriptor 30 out of 52: Middle Aged
[ Info: Descriptor 29 out of 52: Weight Gain
[ Info: Descriptor 33 out of 52: Young Adult
[ Info: Descriptor 34 out of 52: Incidence
[ Info: Descriptor 31 out of 52: Surveys and Questionnaires
[ Info: Descriptor 39 out of 52: Metabolic Syndrome
[ Info: Descriptor 43 out of 52: Sleep Apnea, Obstructive
[ Info: Descriptor 41 out of 52: Nutrition Surveys
[ Info: Descriptor 38 out of 52: Waist Circumference
[ Info: Descriptor 44 out of 52: Television
[ Info: Descriptor 46 out of 52: Overweight
[ Info: Descriptor 37 out of 52: Pain Measurement
[ Info: Descriptor 48 out of 52: Female
[ Info: Descriptor 47 out of 52: Physical Fitness
[ Info: Descriptor 45 out of 52: Age Distribution
[ Info: Descriptor 3 out of 52: ROC Curve
[ Info: Descriptor 42 out of 52: Fibromyalgia
[ Info: Descriptor 2 out of 52: Republic of Korea
[ Info: Descriptor 50 out of 52: Obesity, Abdominal
[ Info: Descriptor 7 out of 52: Logistic Models
[ Info: Descriptor 49 out of 52: Biomarkers
┌ Warning: `getindex(df::DataFrame, col_ind::ColumnIndex)` is deprecated, use `df[!, col_ind]` instead.
│   caller = insert_row!(::SQLite.DB, ::String, ::Dict{Symbol,String}, ::Bool) at sqlite_db_utils.jl:59
└ @ BioMedQuery.DBUtils ~/build/bcbi/BioMedQuery.jl/src/DBUtils/sqlite_db_utils.jl:59
[ Info: Descriptor 51 out of 52: C-Reactive Protein
[ Info: Descriptor 40 out of 52: Pain
[ Info: Descriptor 36 out of 52: Body Weight
[ Info: Descriptor 35 out of 52: Prognosis
[ Info: Descriptor 32 out of 52: Health Status
[ Info: Descriptor 52 out of 52: Aged, 80 and over
[ Info: Descriptor 51 out of 52: C-Reactive Protein
  2.894164 seconds (2.26 M allocations: 110.942 MiB, 2.26% gc time)

Explore the output table

db_query(db_sqlite, "SELECT * FROM mesh2umls;")

56 rows × 2 columns

	mesh	umls
	String⍰	String⍰
1	Reference Values	Quantitative Concept
2	Severity of Illness Index	Quantitative Concept
3	Obesity	Disease or Syndrome
4	Fatigue	Sign or Symptom
5	Sedentary Behavior	Finding
6	Male	Organism Attribute
7	Prevalence	Quantitative Concept
8	Case-Control Studies	Research Activity
9	Risk Factors	Finding
10	Analysis of Variance	Quantitative Concept
11	Odds Ratio	Quantitative Concept
12	Thyrotropin	Amino Acid, Peptide, or Protein
13	Thyrotropin	Hormone
14	Thyrotropin	Pharmacologic Substance
15	Exercise	Daily or Recreational Activity
16	Quality of Life	Idea or Concept
17	Body Mass Index	Diagnostic Procedure
18	Sex Distribution	Quantitative Concept
19	Humans	Human
20	Pediatric Obesity	Disease or Syndrome
21	Germany	Geographic Area
22	Child	Age Group
23	Sports	Daily or Recreational Activity
24	Adult	Age Group
25	Cross-Sectional Studies	Research Activity
26	Middle Aged	Age Group
27	Japan	Geographic Area
28	Aged	Organism Attribute
29	Surveys and Questionnaires	Research Activity
30	Incidence	Quantitative Concept
31	Nutrition Surveys	Research Activity
32	Weight Gain	Finding
33	Sleep Apnea, Obstructive	Disease or Syndrome
34	Physical Fitness	Idea or Concept
35	Pain Measurement	Diagnostic Procedure
36	Young Adult	Age Group
37	Age Distribution	Quantitative Concept
38	Fibromyalgia	Disease or Syndrome
39	Metabolic Syndrome	Disease or Syndrome
40	Television	Manufactured Object
41	Prognosis	Health Care Activity
42	Body Weight	Organism Attribute
43	Logistic Models	Intellectual Product
44	Logistic Models	Quantitative Concept
45	Female	Population Group
46	Waist Circumference	Clinical Attribute
47	Obesity, Abdominal	Finding
48	C-Reactive Protein	Immunologic Factor
49	C-Reactive Protein	Amino Acid, Peptide, or Protein
50	Republic of Korea	Geographic Area
51	Pain	Food
52	Health Status	Qualitative Concept
53	Overweight	Finding
54	Biomarkers	Clinical Attribute
55	ROC Curve	Quantitative Concept
56	Aged, 80 and over	Age Group

Filtering MeSH terms by UMLS concept

Getting the descriptor to index dictionary and occurence matrix

@time labels2ind, occur = umls_semantic_occurrences(db_sqlite, umls_concept);

(Dict("Obesity"=>1,"Pediatric Obesity"=>2,"Sleep Apnea, Obstructive"=>3,"Metabolic Syndrome"=>4,"Fibromyalgia"=>5), 
  [1, 1]  =  1.0
  [5, 1]  =  1.0
  [1, 2]  =  1.0
  [3, 2]  =  1.0
  [2, 3]  =  1.0
  [1, 4]  =  1.0
  [4, 5]  =  1.0)

Descriptor to Index Dictionary

labels2ind

Dict{String,Int64} with 5 entries:
  "Obesity"                  => 1
  "Pediatric Obesity"        => 2
  "Sleep Apnea, Obstructive" => 3
  "Metabolic Syndrome"       => 4
  "Fibromyalgia"             => 5

Output Data Matrix

Matrix(occur)

5×5 Array{Float64,2}:
 1.0  1.0  0.0  1.0  0.0
 0.0  0.0  1.0  0.0  0.0
 0.0  1.0  0.0  0.0  0.0
 0.0  0.0  0.0  0.0  1.0
 1.0  0.0  0.0  0.0  0.0

DataFrames

This example demonstrates the typical workflow to create a MeSH to UMLS map as a DataFrame relating all concepts associated with all MeSH terms in the input dataframe.

Get the articles (same as example in PubMed Search and Parse)

dfs = Processes.pubmed_search_and_parse(email, search_term, max_articles, verbose)

Dict{String,DataFrames.DataFrame} with 8 entries:
  "basic"               => 5×13 DataFrames.DataFrame. Omitted printing of 9 col…
  "mesh_desc"           => 52×2 DataFrames.DataFrame…
  "mesh_qual"           => 9×2 DataFrames.DataFrame…
  "pub_type"            => 10×3 DataFrames.DataFrame…
  "abstract_full"       => 5×2 DataFrames.DataFrame. Omitted printing of 1 colu…
  "author_ref"          => 35×8 DataFrames.DataFrame. Omitted printing of 3 col…
  "mesh_heading"        => 78×5 DataFrames.DataFrame…
  "abstract_structured" => 4×4 DataFrames.DataFrame. Omitted printing of 1 colu…

Map MeSH to UMLS and explore the output table

@time res = map_mesh_to_umls_async(dfs["mesh_desc"], umls_user, umls_pswd)

56 rows × 2 columns

	descriptor	concept
	String	String
1	Adult	Age Group
2	Age Distribution	Quantitative Concept
3	Aged	Organism Attribute
4	Aged, 80 and over	Age Group
5	Analysis of Variance	Quantitative Concept
6	Biomarkers	Clinical Attribute
7	Body Mass Index	Diagnostic Procedure
8	Body Weight	Organism Attribute
9	C-Reactive Protein	Amino Acid, Peptide, or Protein
10	C-Reactive Protein	Immunologic Factor
11	Case-Control Studies	Research Activity
12	Child	Age Group
13	Cross-Sectional Studies	Research Activity
14	Exercise	Daily or Recreational Activity
15	Fatigue	Sign or Symptom
16	Female	Population Group
17	Fibromyalgia	Disease or Syndrome
18	Germany	Geographic Area
19	Health Status	Qualitative Concept
20	Humans	Human
21	Incidence	Quantitative Concept
22	Japan	Geographic Area
23	Logistic Models	Intellectual Product
24	Logistic Models	Quantitative Concept
25	Male	Organism Attribute
26	Metabolic Syndrome	Disease or Syndrome
27	Middle Aged	Age Group
28	Nutrition Surveys	Research Activity
29	Obesity	Disease or Syndrome
30	Obesity, Abdominal	Finding
31	Odds Ratio	Quantitative Concept
32	Overweight	Finding
33	Pain	Food
34	Pain Measurement	Diagnostic Procedure
35	Pediatric Obesity	Disease or Syndrome
36	Physical Fitness	Idea or Concept
37	Prevalence	Quantitative Concept
38	Prognosis	Health Care Activity
39	Quality of Life	Idea or Concept
40	ROC Curve	Quantitative Concept
41	Reference Values	Quantitative Concept
42	Republic of Korea	Geographic Area
43	Risk Factors	Finding
44	Sedentary Behavior	Finding
45	Severity of Illness Index	Quantitative Concept
46	Sex Distribution	Quantitative Concept
47	Sleep Apnea, Obstructive	Disease or Syndrome
48	Sports	Daily or Recreational Activity
49	Surveys and Questionnaires	Research Activity
50	Television	Manufactured Object
51	Thyrotropin	Amino Acid, Peptide, or Protein
52	Thyrotropin	Hormone
53	Thyrotropin	Pharmacologic Substance
54	Waist Circumference	Clinical Attribute
55	Weight Gain	Finding
56	Young Adult	Age Group

Getting the descriptor to index dictionary and occurence matrix

@time labels2ind, occur = umls_semantic_occurrences(dfs, res, umls_concept);

(Dict("Obesity"=>1,"Pediatric Obesity"=>2,"Sleep Apnea, Obstructive"=>3,"Metabolic Syndrome"=>4,"Fibromyalgia"=>5), 
  [4, 1]  =  1.0
  [1, 2]  =  1.0
  [2, 3]  =  1.0
  [1, 4]  =  1.0
  [3, 4]  =  1.0
  [1, 5]  =  1.0
  [5, 5]  =  1.0)

Descriptor to Index Dictionary

labels2ind

Dict{String,Int64} with 5 entries:
  "Obesity"                  => 1
  "Pediatric Obesity"        => 2
  "Sleep Apnea, Obstructive" => 3
  "Metabolic Syndrome"       => 4
  "Fibromyalgia"             => 5

Output Data Matrix

Matrix(occur)

5×5 Array{Float64,2}:
 0.0  1.0  0.0  1.0  1.0
 0.0  0.0  1.0  0.0  0.0
 0.0  0.0  0.0  1.0  0.0
 1.0  0.0  0.0  0.0  0.0
 0.0  0.0  0.0  0.0  1.0

This page was generated using Literate.jl.