Animal Subject Shape
The project defines a set of re-usable, hierarchical shapes for use in a variety of data validation scenarios spanning the study data lifecycle. The first shapes are constructed for the Demographics (DM) domain, where each Animal Subject is a represented as a row in the source data. This makes the AnimalSubjectShape a natural choice for a shape containing additional constraints. The Animal Subject Shape definitions are located in this file:
- AnimalSubjectShape SHACL/CJ16050Constraints/SHACL-AnimalSubject.TTL.
The SHACL shapes structure and naming conventions are specified with reuse in mind. Where practical, shapes are named using a description of their function plus the word Shape followed by a dash and then an abbreviated name of the class or entity they act upon. Examples:
hasMin1Max1Shape-USubjID- validates that each Animal Subject has a minimum of one and maximum of one USUBJID value.isUniqueShape-USubjID- validates the uniqueness of USUBJID values. A USUBJID cannot be assigned to more than one Animal Subject.
Shapes may include additional constraints such as data type, length, and other restrictions not explicitly stated in the original FDA rules.
The sh:message property provides meaningful messages about violations when they are detected. Where applicable, the related FDA Rule ID number is provided in square brackets at the end of the message text. In cases where a shape may be applied to more than one Rule, all rules covered by that shape are listed.
Example: sh:message "Subject --> USUBJID violation. [SD0083]" ;
A shape is created to define the constraints attached to the Animal Subject IRI. Each individual constraint is described in the sections that follow.
sh:property for each type of predicate ----> object relation attached directly to the AnimalSubject IRI.
predicate ----> object relation for the AnimalSubject class, with the exception of predicates like `rdf:type`, `skos:prefLabel`, etc., has a `sh:property` definition for a shape that validates that type of entity.
The Node Shape study:AnimalSubjectShape describes nodes of the class study:AnimalSubject . FDA Rule numbers are added as comments to facilitate referencing back to the original FDA requirements.
study:AnimalSubjectShape a sh:NodeShape ; sh:targetClass study:AnimalSubject ; sh:property study:hasMin1Max1Shape-USubjID ; # Rule SD0083 sh:property study:isUniqueShape-USubjID ; # Rule SD0083 sh:property study:hasMin1Max1Shape-SubjID ; # Rule SD1001 sh:property study:isUniqueShape-SubjID ; # Rule SD1001 sh:property study:hasTypeXsdDate-Date ; # Rule SD1002 sh:property study:hasMin1Max1Shape-Interval ; # Rule SD1002 sh:property study:hasMin1Max1Shape-StartEndDates ; # Rule SD1002 sh:property study:hasStartLEEndShape-Interval ; # Rule SD1002 sh:property study:hasMinInclusive0Shape-Age ; # Rule SD0084 ... more property shapes will be added as they are developed
If an ontology defines study:AnimalSubject as a subclass of study:Subject, then shapes could use the sh:targetClass study:Subject (assuming common constraints for both classes.) If a clinical trial on human study subjects were to define a study:HumanStudySubject as a subclass of study:Subject, the same constraints could be used for both pre-clinical (non-human animals) and clinical (human) data validation. This work focusses on SEND, so the target class study:AnimalSubject is specified for simplicity.
study:Subject rdf:type owl:Class ; rdfs:subClassOf study:Party ; skos:prefLabel "Subject" ; . study:Animal rdf:type owl:Class ; rdfs:subClassOf study:BiologicEntity ; skos:prefLabel "Animal" ; . study:AnimalSubject rdf:type owl:Class ; rdfs:subClassOf study:Animal ; rdfs:subClassOf study:Subject ; skos:prefLabel "Animal subject" ; .
Data
It is important to understand how the data is restructured from row-by-column XPT format by the Data Conversion process. Here is a subset of the original data for Animal Subject 00M01:
| SubjectIRI | subjid | usubjid | setcd | species |
|---|---|---|---|---|
| Animal_a6d09184 | 00M01 | CJ16050_00M01 | 00 | Rat |
And its translation into RDF triples:
cj16050:Animal_a6d09184
a study:AnimalSubject ;
skos:prefLabel "Animal 00M01"^^xsd:string ;
study:hasUniqueSubjectID cj16050:UniqueSubjectIdentifier_CJ16050_00M01 ;
study:hasSubjectID cj16050:SubjectIdentifier_00M01 ;
study:hasReferenceInterval cj16050:Interval_Animal_a6d09184 ;
study:memberOf cjprot:Set_00,
code:Species_Rat ;
study:participatesIn cj16050:SexDataCollection_Animal_a6d09184 ,
cj16050:AgeDataCollection_Animal_a6d09184,
cj16050:Randomization_Animal_a6d09184 .
The first validation shapes will be formed around the AnimalSubject Class study:AnimalSubject as described above. The first rules in our project center around the subject identifiers usubjid and subjid. The figure below shows the connections from the Animal Subject IRI to the USUBJID and SUBJID IRI values. Become familiar with both the figure and the RDF triples (above) before proceeding to the next steps.
Animal Subject Node to ID Values
FDA Rules as SHACL Shapes
Rule SD0083 : USUBJID
The spreadsheet FDA-Validator-Rules.xlsx defines the rule for USUBJID in the DM Domain as:
| FDA Validator Rule ID | FDA Validator Message | Business or Conformance Rule Validated | FDA Validator Rule |
|---|---|---|---|
| SD0083 | Duplicate USUBJID | Identifier used to uniquely identify a subject across all studies | The value of Unique Subject Identifier (USUBJID) variable must be unique for each subject across all trials* in the submission. |
* Because the prototype is based on data from a single trial, Rule SD0083 is evaluated within the context of a single study.
The Rule is deconstructed into the following components based on knowledge of the study data requirements, RDF data model (schema), and SD0083 rule statement:
RC1. An Animal Subject cannot have more than one USUBJID.
RC2. An Animal Subject cannot have a missing USUBJID.
RC3. A USUBJID cannot be assigned to more than one Animal Subject.
Translation of Rule Components into SHACL and evaluation of test data is described below. Rule Components RC1 and RC2 are satisfied by a single SHACL Shape, while a second shape evaluates the third component. Addition information for the traceability between rules, shapes, and the data used to evaluate them is available in the file TestCases.xlsx
SD0083-RC1, RC2 : A single, non-missing USUBJID per Animal Subject.
A copy of the DM TTL file for SHACL testing is available at SHACL/CJ16050Constraints/DM-CJ16050-R.TTL
| studyid | domain | usubjid | subjid | SubjectIRI | Rule Violated |
|---|---|---|---|---|---|
| CJ16050 | DM | CJ16050_00M01 | 00M01 | Animal_a6d09184 | None |
| CJ16050 | DM | CJ16050_99T1 | 99T1 | Animal_2a836191 | SD0083-RC1 |
| CJ16050 | DM | CJ16050_99T2 | 99T2 | Animal_2a836191 | SD0083-RC1 |
| CJ16050 | DM | NA | NA | Animal_69fa85ac | SD0083-RC2 |
Referring to the AnimalSubject IRIs (column SubjectIRI):
- Animal_a6d09184 is compliant, with single values for
usubjid,subjid. - Animal_2a836191 has two
usubjidvalues and twosubjidvalues. - Animal_69fa85ac is missing values for both
usubjidandsubjid
Triples for the compliant AnimalSubject appear as:
cj16050:Animal_a6d09184
a study:AnimalSubject ;
skos:prefLabel "Animal 00M01"^^xsd:string ;
study:hasUniqueSubjectID cj16050:UniqueSubjectIdentifier_CJ16050_00M01 ;
...
Each AnimalSubject must be evaluated to ensure it has minimum of one usubjid and no more than one usubjid. The shape will therefore be named study:hasMin1Max1Shape-USubjID and it is a sh:property of the study:AnimalSubjectShape.
study:AnimalSubjectShape a sh:NodeShape ; sh:targetClass study:AnimalSubject sh:property study:hasMin1Max1Shape-USubjID ; ...
From the TTL data we see that usubjid is attached to the AnimalSubject IRI by the study:hasUniqueSubjectID predicate. This predicate is specified as the object (target) of sh:path as shown in the shape definition, below. The shape further specifies there should be a minimum of one and maximum of one path through study:hasUniqueSubjectID for a given AnimalSubject.
study:AnimalSubjectShape a sh:NodeShape ; sh:targetClass study:AnimalSubject sh:property study:hasMin1Max1Shape-USubjID ; ... #--- Unique Subject ID (USUBJID) ---- study:hasMin1Max1Shape-USubjID a sh:PropertyShape ; sh:name "minmaxUniqueSubjid" ; sh:description "A single, exclusive USUBJID must be assigned to a Subject." ; sh:message "Subject --> USUBJID violation. [SD0083]" ; sh:path study:hasUniqueSubjectID ; sh:minCount 1 ; sh:maxCount 1 .
SHACL file is located at /SENDConform/SHACL/CJ16050Constraints/SHACL-AnimalSubject.TTL
SD0083-Test Case 1 : Animal Subject Assigned Two USUBJID values
Test data for Animal Subject IRI Animal_2a836191 is assigned to two USUBJID values, violating SD0083-RC1.
cj16050:Animal_2a836191
a study:AnimalSubject ;
skos:prefLabel "Animal 99T1"^^xsd:string,
"Animal 99T2"^^xsd:string ;
study:hasUniqueSubjectID cj16050:UniqueSubjectIdentifier_CJ16050_99T1,
cj16050:UniqueSubjectIdentifier_CJ16050_99T2 ;
...
Violation of Rule Component 1 as detected by the sh:maxCount constraint:
... sh:path study:hasUniqueSubjectID ; sh:minCount 1 ; sh:maxCount 1 ...
See the instructions on the Validation page for how to apply the SHACL to the data and generate the Validation Report.
The Report correctly confirms AnimalSubject Animal_2a836191 has more than one USUBJID value, violating the MaxConstraintComponent of FDA Rule SD0083.
a sh:ValidationResult ;
sh:resultSeverity sh:Violation ;
sh:sourceShape study:hasMin1Max1Shape-USubjID ;
sh:focusNode cj16050:Animal_2a836191 ;
sh:resultMessage "Subject --> USUBJID violation. [SD0083]" ;
sh:resultPath study:hasUniqueSubjectID ;
sh:sourceConstraintComponent sh:MaxCountConstraintComponent
The AnimalSubject IRI in the Report can be use to identify the USUBJID value that violates the constraint. File: /SPARQL/SD0083-TC1-Info.rq
SELECT ?animalIRI ?usubjidLabel
WHERE{
cj16050:Animal_2a836191 study:hasUniqueSubjectID ?usubjidIRI .
?usubjidIRI skos:prefLabel ?usubjidLabel .
BIND(IRI(cj16050:Animal_2a836191) AS ?animalIRI )
}
The query result shows Animal_2a836191 is assigned two usubjid, in violation of the rule.
animalIRI usubjidLabel cj16050:Animal_2a836191 "CJ16050-99T1" cj16050:Animal_2a836191 "CJ16050_99T2"Verify
SPARQL independently verifies Animal_2a836191 has two USUBJID values by querying the graph to find all Subject IRIs that do not have one and only 1 usubjidIRI. File: /SPARQL/SD0083-TC1-Verify.rq
SELECT ?animalSubjectIRI (COUNT(?usubjidIRI) AS ?total)
WHERE{
?animalSubjectIRI a study:AnimalSubject ;
study:hasUniqueSubjectID ?usubjidIRI .
?usubjidIRI skos:prefLabel ?usubjidLabel .
} GROUP BY ?animalSubjectIRI
HAVING (?total != 1)
animalSubjectIRI total cj16050:Animal_2a836191 2
In the test data, animalSubject IRI Animal_69fa85ac has no USUBJID value, violating SD0083-RC2.
cj16050:Animal_69fa85ac
a study:AnimalSubject ;
study:hasReferenceInterval cj16050:Interval_Animal_69fa85ac ;
study:memberOf cjprot:Set_00, code:Species_Rat ;
study:participatesIn cj16050:AgeDataCollection_Animal_69fa85ac, cj16050:SexDataCollection_Animal_69fa85ac .
The <a href=’#sd0083rc1shacl>SHACL is identical to SD0083-Test Case 1</a>.
The Report correctly identifies AnimalSubject IRI Animal_69fa85ac as violating the constraint that USUBJID must be non-missing.
a sh:ValidationResult ;
sh:resultSeverity sh:Violation ;
sh:sourceShape study:hasMin1Max1Shape-USubjID ;
sh:focusNode cj16050:Animal_69fa85ac ;
sh:resultMessage "Subject --> USUBJID violation [SD0083]" ;
sh:resultPath study:hasUniqueSubjectID ;
sh:sourceConstraintComponent sh:MinCountConstraintComponent
The AnimalSubject IRI in the Report can be use to identify the value of Predicates and Objects attached to the AnimalSubject IRI in facilitate identification of the problematic record, since a missing USUBJID means no skos:prefLabel is available. File: /SPARQL/SD0083-TC2-Info.rq
SELECT ?animalIRI ?p ?o
WHERE{
cj16050:Animal_69fa85ac ?p ?o .
BIND(IRI(cj16050:Animal_Animal_69fa85ac) AS ?animalIRI )
}
animalIRI p o cj16050:Animal_69fa85ac rdf:type study:AnimalSubject cj16050:Animal_69fa85ac study:hasReferenceInterval cj16050:Interval_Animal_69fa85ac cj16050:Animal_69fa85ac study:memberOf cjprot:#Set_00 cj16050:Animal_69fa85ac study:memberOf code:Species_Rat cj16050:Animal_69fa85ac study:participatesIn cj16050:AgeDataCollection_Animal_69fa85ac cj16050:Animal_69fa85ac study:participatesIn cj16050:SexDataCollection_Animal_69fa85acVerify
SPARQL independently confirms Animal_69fa85acc has no USUBJID. Because usubjid is used as the skos:prefLabel for AnimalSubject, there is not label to return when usubjid is missing. File: /SPARQL/SD0083-TC2-Verify.rq
SELECT ?animalIRI
WHERE{
?animalIRI a study:AnimalSubject .
OPTIONAL{ ?animalIRI study:hasUniqueSubjectID ?usubjid . }
FILTER(NOT EXISTS { ?animalIRI study:hasUniqueSubjectID ?usubjid. })
}
animalIRI cj16050:Animal_69fa85ac
SD0083-RC3: A USUBJID cannot be assigned to more than one Animal Subject Test Data
In the test data, Animal Subjects Animal_5dba5b4b and Animal_1a2751f1 have the same USUBJID value.
| studyid | domain | usubjid | subjid | SubjectIRI | Rule Violated |
|---|---|---|---|---|---|
| CJ16050 | DM | CJ16050_00M01 | 00M01 | Animal_a6d09184 | None |
| CJ16050 | DM | CJ16050_99T4 | 99T4 | Animal_5dba5b4b | SD0083-RC3 |
| CJ16050 | DM | CJ16050_99T4 | 99T4 | Animal_1a2751f1 | SD0083-RC3 |
In RDF:
cj16050:Animal_5dba5b4b
a study:AnimalSubject ;
skos:prefLabel "Animal 99T4"^^xsd:string ;
study:hasUniqueSubjectID cj16050:UniqueSubjectIdentifier_CJ16050_99T4 ;
cj16050:Animal_1a2751f1
a study:AnimalSubject ;
skos:prefLabel "Animal 99T4"^^xsd:string ;
study:hasUniqueSubjectID cj16050:UniqueSubjectIdentifier_CJ16050_99T4 ;
There are multiple ways to assess the unique assignment of a USUBJID value to an AnimalSubject. Both SHACL-Core and SHACL-SPARQL are viable alternatives. Two SHACL-Core alternatives are discussed here.
Method 1: Identify USUBJIDs assigned to multiple AnimalSubjects
- Identify duplicated USUBJID values.
SHACL Shape for Method 1: Identify duplicate USUBJID values. This shape is applied to all uses of the predicate study:hasUniqueSubjectID, allowing its use for both SEND and SDTM when that predicate is present.
study:isUniqueShape-USubjID a sh:PropertyShape ;
sh:targetObjectsOf study:hasUniqueSubjectID ;
sh:name "uniqueUSubjid" ;
sh:description "A USUBJID must only be assigned to one Subject." ;
sh:message "USUBJID assigned to more than one Subject. [SD0083]" ;
sh:property [
sh:path [sh:inversePath study:hasUniqueSubjectID] ;
sh:maxCount 1
] .
The Validation Report from Method 1 is not shown because Method 2 was chosen for the project.
Method 2: Identify the AnimalSubjects that have the same USUBJID
- Identify AnimalSubject IRIs assigned to the same USUBJID value.
# Animal Subject Shape
study:AnimalSubjectShape
a sh:NodeShape ;
sh:targetClass study:AnimalSubject
sh:property study:hasMin1Max1Shape-USubjID ;
sh:property study:isUniqueShape-USubjID
...
study:isUniqueShape-USubjID
a sh:PropertyShape ;
sh:name "uniqueUSubjid" ;
sh:description "A USUBJID must only be assigned to one Subject." ;
sh:message "USUBJID assigned to more than one Subject. [SD0083]" ;
sh:path (study:hasUniqueSubjectID [sh:inversePath study:hasUniqueSubjectID]) ;
sh:maxCount 1 ;
.
Method 2 was chosen for consistency with the other checks in this section that focus on the identification of AnimalSubjects that fail constraints.
The report from Method 2 correctly identifies the Animal Subjects Animal_5dba5b4b and Animal_1a2751f1 as sharing the same USUBJID.
a sh:ValidationResult ;
sh:sourceConstraintComponent sh:MaxCountConstraintComponent ;
sh:focusNode cj16050:Animal_5dba5b4b ;
sh:sourceShape _:bnode_dc2d5e41_a650_456a_87ee_944f84cffae6_826 ;
sh:resultPath ( study:hasUniqueSubjectID [
sh:inversePath study:hasUniqueSubjectID
] ) ;
sh:resultMessage "USUBJID assigned to more than one Subject. [SD0083]" ;
sh:resultSeverity sh:Violation
...
a sh:ValidationResult ;
sh:sourceConstraintComponent sh:MaxCountConstraintComponent ;
sh:focusNode cj16050:Animal_1a2751f1 ;
sh:sourceShape _:bnode_dc2d5e41_a650_456a_87ee_944f84cffae6_826 ;
sh:resultPath ( study:hasUniqueSubjectID [
sh:inversePath study:hasUniqueSubjectID
] ) ;
sh:resultMessage "USUBJID assigned to more than one Subject. [SD0083]" ;
sh:resultSeverity sh:Violation ;
...
Use the AnimalSubject IRI values to identify the offendingusubjid value. File: /SPARQL/SD0083-TC3-Info.rq
SELECT ?animalIRI ?animalLabel ?usubjid
WHERE{
{
cj16050:Animal_5dba5b4b study:hasUniqueSubjectID ?usubjidIRI ;
skos:prefLabel ?animalLabel .
?usubjidIRI skos:prefLabel ?usubjid .
BIND(IRI(cj16050:Animal_5dba5b4b) AS ?animalIRI )
}
UNION
{
cj16050:Animal_1a2751f1 study:hasUniqueSubjectID ?usubjidIRI ;
skos:prefLabel ?animalLabel .
?usubjidIRI skos:prefLabel ?usubjid .
BIND(IRI(cj16050:Animal_1a2751f1) AS ?animalIRI )
}
}
animalIRI animalLabel usubjid cj16050:Animal_5dba5b4b "Animal 99T4" "CJ16050_99T4" cj16050:Animal_1a2751f1 "Animal 99T4" "CJ16050_99T4"Verify
Independently verify Animal_5dba5b4b and Animal_1a2751f1 share the same USUBJID (and consequently the same label for the AnimalSubject and USUBJID). File: /SPARQL/SD0083-TC3-Verify.rq
SELECT ?animalIRI ?usubjid
WHERE{
?animalIRI study:hasUniqueSubjectID ?usubjidIRI ;
skos:prefLabel ?usubjid .
?animalIRI2 study:hasUniqueSubjectID ?usubjidIRI .
FILTER(?animalIRI != ?animalIRI2)
}
animalIRI usubjid cj16050:Animal_1a2751f1 "Animal 99T4" cj16050:Animal_5dba5b4b "Animal 99T4"
The following information is outdated as of 2020-02-13.
Rule SD1001 : SUBJID
The spreadsheet FDA-Validator-Rules.xlsx defines the rule for SUBJID in the DM Domain as:
| FDA Validator Rule ID | FDA Validator Message | Business or Conformance Rule Validated | FDA Validator Rule |
|---|---|---|---|
| SD1001 | Duplicate SUBJID | ‘Subject identifier, which must be unique within the study. | The value of Subject Identifier for the Study (SUBJID) variable must be unique for each subject within the study. |
The Rule Components and corresponding SHACL shapes for SD1001 are similar to those defined for USUBJID in SD0083 with exception of the predicate changing to study:hasSubjectID and result messages specific to SUBJID instead of USUBJID. Details for SD1001 are therefore not provided here. The SHACL is available in the file SHACL-AnimalSubject.TTL
Rule SD1002 : RFSTDTC, RFENDTC (Reference Interval)
The spreadsheet FDA-Validator-Rules.xlsx defines Rule SD10002 for Reference Start Date (RFSTDTC) and Reference End Date (RFENDTC) as:
| FDA Validator Rule ID | FDA Validator Message | Business or Conformance Rule Validated | FDA Validator Rule |
|---|---|---|---|
| SD1002 | RFSTDTC is after RFENDTC | Study Start and End Dates must be submitted and complete. | Subject Reference Start Date/Time (RFSTDTC) must be less than or equal to Subject Reference End Date/Time (RFENDTC) |
This project models RFSTDTC and RFENDTC as part of a Reference Interval (see next figure) connected to the Animal Subject IRI. This leads to the deconstruction of the FDA rule into the following Rule Components:
RC1. Reference Start Date and End Date must be in xsd:date format.
RC2. An Animal Subject has one Reference Interval.
RC3. A Reference Interval has one Start Date and one End Date.
RC4. Start Date must be on or before End Date.
Data StructureFamiliarity with the data structure is necessary to explain the constraints and test cases. The figure below illustrates a partial set of data for an AnimalSubject where the Reference Interval end date precedes the start date, thus violating Rule Component 4 of SD1002.
AnimalSubject with incorrect Reference Interval dates
SD1002-RC1. Reference Start Date and End Date in xsd:date formatrfstdtc and rfendtc in xsd:date format.
xsd:date. Other studies may use xsd:dateTime or a combination of xsd:date
and xsd:dateTime.
In the test data, two Animal Subjects have string values for dates instead of the required xsd:date.
| usubjid | SubjectIRI | rfstdtc | rfendtc | Rule Violated |
|---|---|---|---|---|
| 00M01 | Animal_a6d09184 | 2016-12-07 | 2016-12-07 | None |
| 99T6 | Animal_aa573a5d | 5-DEC-16 | 2016-12-07 | SD1002-RC1 |
| 99T7 | Animal_cdd31fb6 | 2016-12-07 | 6-DEC-16 | SD1002-RC1 |
OUTDATED TEXT BEGINS Here
Refer back to previous sections to compare the data to the SHACL, below. The shape :DateFmtShape uses sh:targetObjectsOf to begin evaluation at the object of the predicates time:hasBeginning and time:hasEnd. These objects must be of type study:ReferenceBegin or study:ReferenceEnd and have the predicate time:inXSDDate that leads to the date value that must be in xsd:date format.
Interval IRI - - - time:hasBeginning - - > Date IRI - - > time:inXSDDate - - > Date value Interval IRI - - - time:hasEnd - - > Date IRI - - > time:inXSDDate - - > Date value
Test data for Animal Subject 99T4 contains a string value for rfendtc. Not shown: Subject 9T10 with string value for rfstdtc.
cj16050:Animal_68bab561
a study:AnimalSubject ;
skos:prefLabel "Animal 99T4"^^xsd:string ;
study:hasReferenceInterval cj16050:Interval_Animal_68bab561 ;
...
cj16050:Interval_Animal_68bab561
a study:ReferenceInterval ;
time:hasBeginning cj16050:Date_2016-12-08 ;
time:hasEnd cj16050:Date_7-DEC-16 .
cj16050:Date_7-DEC-16
a study:ReferenceEnd ;
time:inXSDDate "7-DEC-16"^^xsd:string .
The shape tests the following conditions:
- A Reference Start Date must be in
xsd:dateformat. - A Reference End Date must be in
xsd:dateformat.
Additional dates can be assessed by adding additional predicates as sh:targetObjectsOf if the date follows through the path time:inXSDDate.
study:hasTypeXsdDateShape-Date a sh:NodeShape ;
sh:targetObjectsOf time:hasBeginning ;
sh:targetObjectsOf time:hasEnd ;
sh:or (
[ sh:class study:ReferenceBegin ]
[ sh:class study:ReferenceEnd ]
) ;
] ;
sh:property [
sh:path time:inXSDDate ;
sh:datatype xsd:date ;
sh:name "xsd:date format";
sh:description "Date format as xsd:date.";
sh:message "Date not in xsd:date format. [SD1002]"
] .
The report correctly identifies the value ‘7-DEC-16’ as a string, violating the xsd:date requirement.
a sh:ValidationReport ;
sh:conforms false ;
sh:result [
a sh:ValidationResult ;
sh:resultPath time:inXSDDate ;
sh:resultSeverity sh:Violation ;
sh:resultMessage "Date not in xsd:date format. [SD1002]" ;
sh:value "7-DEC-16" ;
sh:sourceShape _:bnode_3c9cf811_13d4_43cb_b212_b7097d00b1ed_221 ;
sh:sourceConstraintComponent sh:DatatypeConstraintComponent ;
sh:focusNode cj16050:Date_7-DEC-16 ;
]
The Report identifies the dates “7-DEC-16” and “6-DEC-16” (not shown above). Execute the following SPARQL to find corresponding Animal SUBJECT IRIs and values (Animal 99T4 for date “7-Dec-16” and Animal 99T10 for date “6-Dec-16”). Source file: /SPARQL/Animal-RefInterval.rq
# RC1 : Find Subject with incorrect date format
SELECT ?animalSubjectIRI ?animalLabel ?date
WHERE{
?animalSubjectIRI a study:AnimalSubject ;
skos:prefLabel ?animalLabel ;
study:hasReferenceInterval ?intervalIRI .
?intervalIRI ?beginOrEnd ?dateIRI .
?dateIRI time:inXSDDate ?date .
FILTER (?dateIRI IN (cj16050:Date_6-DEC-16, cj16050:Date_7-DEC-16))
}
Verify
SPARQL independently verifies the test case by finding the two dates that are incorrectly typed as strings. Source file: /SPARQL/Animal-RefInterval.rq
# RC1 : Independently verify with SPARQL
SELECT ?refIntervalIRI ?dateIRI ?date ?dateDType
WHERE{
?refIntervalIRI a study:ReferenceInterval ;
?beginOrEnd ?dateIRI .
?dateIRI time:inXSDDate ?date .
FILTER (datatype(?date) != xsd:date)
}
:AnimalSubject :hasReferenceInterval with sh:minCount and sh:maxCount of 1
In the test data, subject 99T8 has no Reference Interval. While is possible to have a reference interval without start and end dates (see Data Conversion, the 99T8 has no interval created during the data conversion process (see RDF, below) specifically to test this constraint. Subject 99T9 has two reference intervals.
| usubjid | SubjectIRI | rfstdtc | rfendtc | Rule Violated |
|---|---|---|---|---|
| 00M01 | Animal_a6d09184 | 2016-12-07 | 2016-12-07 | None |
| 99T8 | Animal_d9209e97 | NA | NA | SD1002-RC2-TC1 |
| 99T9 | Animal_cdd31fb6 | 2016-11-11 | 2016-11-25</font> | SD1002-RC2-TC2 |
| 99T9 | Animal_cdd31fb6 | 2016-12-20 | 2016-12-28</font> | SD1002-RC2-TC2 |
OUTDATED TEXT BEGINS Here
This check determines if the Animal Subject has one and only one Reference Interval IRI. While it is possible to have an Interval IRI with no start date and no end date (see Data Conversion), this rule component only evaluates the case of missing Reference Interval IRIs. Multiple start/end dates for a single subject are evaluated in Rule Component 3.
Test data for Animal Subject 99T11 has no study:hasReferenceInterval .
AnimalSubject with more than one Reference Interval.
cj16050:Animal_2a836191
a study:AnimalSubject ;
skos:prefLabel "Animal 99T11"^^xsd:string ;
study:hasSubjectID cj16050:SubjectIdentifier_6204e90c ;
study:hasUniqueSubjectID cj16050:UniqueSubjectIdentifier_6204e90c ;
study:memberOf cjprot:Set_00, code:Species_Rat ;
study:participatesIn cj16050:AgeDataCollection_Animal_6204e90c, cj16050:SexDataCollection_Animal_6204e90c .
The study ontology definesstudy:AnimalSubject as a sub class of both study:Subject and study:Animal. Study subjects, be they animal or person, have a Reference Interval documenting their participation in a trial. Therefore, when the ontology is loaded into the database, the same constraint can be used for both pre-clinical (SEND) and clinical (SDTM) studies. This same ontological approach is taken for USUBJID and SUBJID.
study:Subject rdf:type owl:Class ; rdfs:subClassOf study:Party ; skos:prefLabel "Subject" ; . study:Animal rdf:type owl:Class ; rdfs:subClassOf study:BiologicEntity ; skos:prefLabel "Animal" ; . study:AnimalSubject rdf:type owl:Class ; rdfs:subClassOf study:Animal ; rdfs:subClassOf study:Subject ; skos:prefLabel "Animal subject" ; .
The SHACL shape evaluates the path study:hasReferenceInterval from the targetClass to determine if one and only one Reference Interval IRI is present. When the ontology is loaded, the more general study:Subject can be leveraged as the targetClass, assuming other study:Subjects like study:HumanStudySubject use the same predicate. The commented-out alternative is also provided for when the ontology is not loaded, or for cases where the constraint should only apply to study:AnimalSubject and not other classes like study:HumanSubject.
study:hasMin1Max1Shape-Interval a sh:NodeShape ; sh:targetClass study:Subject ; # Ontology # sh:targetClass study:AnimalSubject ; # No Ontology sh:path study:hasReferenceInterval ; sh:name "reference interval present"; sh:description "Animal Subject must have one and only one reference interval IRI."; sh:message "Animal Subject does not have one Reference Interval IRI. [SD1002]" ; sh:minCount 1 ; sh:maxCount 1 .
The report identifies the IRI cj16050:Animal_2a836191 , corresponding to Animal Subject 99T11.
a sh:ValidationReport ;
sh:conforms false ;
sh:result [
a sh:ValidationResult ;
sh:sourceShape :SubjectOneRefIntervalShape ;
sh:resultPath study:hasReferenceInterval ;
sh:resultSeverity sh:Violation ;
sh:focusNode cj16050:Animal_2a836191 ;
sh:resultMessage "Animal Subject does not have one Reference Interval IRI. [SD1002]" ;
sh:sourceConstraintComponent sh:MinCountConstraintComponent
]
SPARQL identifies the reported IRI as belonging to AnimalSubject 99T11, also confirming there is no study:hasReferenceInterval predicate.
# RC 2 : Information : predicates and objects for the IRI in the report
SELECT ?s ?p ?o
WHERE {
cj16050:Animal_2a836191 ?p ?o ;
BIND( IRI(cj16050:Animal_2a836191) as ?s)
} ORDER BY ?p
Verify
Verification identifies Animal Subject 99T11 with no Reference Interval.
# RC 2 : Verify: Number of reference intervals per subject
SELECT ?animalSubjectIRI ?animalLabel (COUNT(?intervalIRI) AS ?numIntervals )
WHERE{
?animalSubjectIRI a study:AnimalSubject ;
skos:prefLabel ?animalLabel ;
OPTIONAL{
?animalSubjectIRI study:hasReferenceInterval ?intervalIRI .
}
} # ORDER BY ?animalLabel
GROUP BY ?animalSubjectIRI ?animalLabel
HAVING (?numIntervals != 1 )
study:ReferenceInterval time:hasBeginning with sh:minCount and sh:maxCount of 1, sh:and time:hasEnd with sh:minCount and sh:maxCount of 1
In the test data, subject 99T8 has no Reference Interval. While is possible to have a reference interval without start and end dates (see Data Conversion, the 99T8 has no interval created during the data conversion process (see RDF, below) specifically to test this constraint. Subject 99T9 has two reference intervals.
| usubjid | SubjectIRI | rfstdtc | rfendtc | Rule Violated |
|---|---|---|---|---|
| 00M01 | Animal_a6d09184 | 2016-12-07 | 2016-12-07 | None |
| 99T11 | Animal_c5e105c3 | NA | 2016-12-07 | SD1002-RC3-TC1 |
| 99T12 | Animal_664e018b | 2016-12-07 | NA | SD1002-RC3-TC2 |
| 99T8 | Animal_d9209e97 | NA | NA | SD1002-RC3-TC3 |
| 99T13 | Animal_8196e3ec | 2016-12-28 | 2016-12-25</font> | SD1002-RC3-TC4 |
Reference interval IRIs are connected to their date values through the paths time:hasBeginning and time:hasEnd. A correctly formed interval has both start and end dates.
Test data provides the following violations:
- 99T11 missing rfstdtc TC 1
- 99T12 missing rfendtc TC 2
- 99T8 missing both rfendtc, rfstdtc TC3
- 99T13 >1 rfstdtc, >1 rfendtc TC4
Only the data and report for 99T5 is shown here, where start date is present and end date is missing for the Reference Interval.
cj16050:Animal_db3c6403 a study:AnimalSubject ; skos:prefLabel "Animal 99T5"^^xsd:string ; study:hasReferenceInterval cj16050:Interval_Animal_db3c6403 ; study:hasSubjectID cj16050:SubjectIdentifier_db3c6403 ; study:hasUniqueSubjectID cj16050:UniqueSubjectIdentifier_db3c6403 ; study:memberOf cjprot:Set_00, code:Species_Rat ; study:participatesIn cj16050:AgeDataCollection_Animal_db3c6403, cj16050:SexDataCollection_Animal_db3c6403 . cj16050:Interval_Animal_db3c6403 a study:ReferenceInterval ; skos:prefLabel "Interval 2016-12-07 NA"^^xsd:string ; time:hasBeginning cj16050:Date_2016-12-07 .
The study ontology definesstudy:ReferenceInterval as a sub class of study:EntityInterval.
study:EntityInterval rdf:type owl:Class ; rdfs:subClassOf time:Interval ; skos:prefLabel "Entity interval" ; . study:ReferenceInterval rdf:type owl:Class ; rdfs:subClassOf study:EntityInterval ; skos:prefLabel "Reference interval" ; . study:Lifespan rdf:type owl:Class ; rdfs:subClassOf study:EntityInterval ; skos:prefLabel "Lifespan" ; . study:MedicalConditionInterval rdf:type owl:Class ; rdfs:subClassOf study:EntityInterval ; skos:prefLabel "Medical event interval" ; . study:StudyParticipationInterval rdf:type owl:Class ; rdfs:subClassOf study:EntityInterval ; skos:prefLabel "Study participation interval" ; .
All sub classes of study:EntityInterval must have a time:hasBeginning and time:hasEnd, allowing the use of a single shape to evaluate following types of intervals when the ontology is loaded into the database:
- study:ReferenceInterval
- study:LifeSpan
- study:MedicalConditionalInterval
- study:StudyParticipationInterval
The ontology facilitates the use of the shape in both pre-clinical (SEND) and clinical (SDTM) studies.
The shape tests the following conditions:
- Interval Start date for an Animal Subject has one and only one value.
- Interval End date for an Animal Subject has one and only one value.
study:hasMin1Max1Shape-StartEndDates a sh:NodeShape ;
sh:targetClass study:EntityInterval ; # Ontology
# sh:targetClass study:ReferenceInterval ; # No Ontology
sh:name "intervalDateCount" ;
sh:description "Interval has one and only one start and end date." ;
sh:message "Problem with Interval date. [SD1002]" ;
sh:and (
[ sh:path time:hasBeginning ;
sh:minCount 1;
sh:maxCount 1
]
[
sh:path time:hasEnd ;
sh:minCount 1;
sh:maxCount 1
]
)
.
The report identifies the interval for Animal Subject 99T5 (cj16050:Interval_Animal_db3c6403) as violating the constraint.
a sh:ValidationResult ;
sh:sourceConstraintComponent sh:AndConstraintComponent ;
sh:focusNode cj16050:Interval_Animal_db3c6403 ;
sh:resultMessage "Problem with Interval date. [SD1002]" ;
sh:value cj16050:Interval_Animal_db3c6403 ;
sh:sourceShape :RefIntervalDateShape ;
sh:resultSeverity sh:Violation ;
SPARQL can trace the reference interval from the report back to AnimalSubject 99T5, showing this individual is missing rfendtc.
SELECT ?animalLabel ?beginDate ?endDate
WHERE{
?animalSubjectIRI study:hasReferenceInterval cj16050:Interval_Animal_db3c6403 ;
skos:prefLabel ?animalLabel .
OPTIONAL{
cj16050:Interval_Animal_db3c6403 time:hasBeginning ?beginIRI .
?beginIRI time:inXSDDate ?beginDate .
}
OPTIONAL{
cj16050:Interval_Animal_db3c6403 time:hasEnd ?endIRI .
?beginIRI time:inXSDDate ?beginDate .
}
}
Verify
The query below correctly lists the AnimalSubjects with start and end date data issues as 99T2, 99T5, 99T8, 99T9.
#--- RC 3: Verify : Pull all subject IDs that do not have one start and one End date
SELECT ?animalLabel ?beginDate ?endDate (COUNT(?beginDate) AS ?numBeginDate)
(COUNT(?endDate) AS ?numEndDate)
WHERE{
?animalSubjectIRI study:hasReferenceInterval ?intervalIRI ;
skos:prefLabel ?animalLabel .
OPTIONAL{
?intervalIRI time:hasBeginning ?beginIRI .
?beginIRI time:inXSDDate ?beginDate .
}
OPTIONAL{
?intervalIRI time:hasEnd ?endIRI .
?beginIRI time:inXSDDate ?endDate .
}
} GROUP BY ?animalSubjectIRI ?animalLabel ?beginDate ?endDate
HAVING ((?numBeginDate != 1) || (?numEndDate != 1) )
Rule Component 4. Start Date on or before End Date
! (?endDate >= ?beginDate )
Referring back to previous sections, the reference start and end dates are not directly attached to either an Animal Subject or that Subject’s Reference Interval IRI. This indirect connection makes the comparison of the two date values more complex, so SHACL-SPARQL is used in place of SHACL-Core. The SPARQL query is written to find cases where the end date is NOT greater than or equal to the start date. Test data provides the following violations:
- 99T1 start date is after end date
- 99T2 multiple start/end date values, one start date is before one end date value
- 99T10 String value for rfstdtc results in a violation when comparing to rfendtc
Only the data and report for 99T1 is shown below.
cj16050:Animal_184f16eb
a study:AnimalSubject ;
skos:prefLabel "Animal 99T1"^^xsd:string ;
study:hasReferenceInterval cj16050:Interval_Animal_184f16eb ;
study:hasSubjectID cj16050:SubjectIdentifier_184f16eb ;
study:hasUniqueSubjectID cj16050:UniqueSubjectIdentifier_184f16eb ;
study:memberOf cjprot:Set_00, code:Species_Rat ;
study:participatesIn cj16050:AgeDataCollection_Animal_184f16eb, cj16050:SexDataCollection_Animal_184f16eb .
cj16050:Interval_Animal_184f16eb
a study:ReferenceInterval ;
skos:prefLabel "Interval 2016-12-07 2016-12-06"^^xsd:string ;
time:hasBeginning cj16050:Date_2016-12-07 ;
time:hasEnd cj16050:Date_2016-12-06 .
cj16050:Date_2016-12-07
a study:ReferenceBegin ;
skos:prefLabel "Date 2016-12-07"^^xsd:string ;
time:inXSDDate "2016-12-07"^^xsd:date ;
study:dateTimeInXSDString "2016-12-07"^^xsd:string .
cj16050:Date_2016-12-06
a study:ReferenceEnd ;
skos:prefLabel "Date 2016-12-06"^^xsd:string ;
time:inXSDDate "2016-12-06"^^xsd:date ;
study:dateTimeInXSDString "2016-12-06"^^xsd:string .
The shape tests the following condition:
- Reference Interval Start Date must be on or before End Date
- The shape will also pick up cases where a date is in
xsd:stringformat.
As described for Rule Component 3, this shape is applied at sh:targetClass study:EntityInterval for intervals that contain time:hasBeginning and time:hasEnd predicates. The alternative application to study:ReferenceInterval is shown for the alternative case when the ontology is not present.
study:hasStartLEEndShape-Interval a sh:NodeShape ;
sh:targetClass study:EntityInterval ; # Ontology
# sh:targetClass study:ReferenceInterval ; # No Ontology
sh:sparql [
a sh:SPARQLConstraint ;
sh:name "sd1002" ;
sh:description "Interval start date on or before end date." ;
sh:message "Interval Start Date on or before End Date";
sh:prefixes [
sh:declare [ sh:prefix "time" ;
sh:namespace "http://www.w3.org/2006/time#"^^xsd:anyURI ;
],
[ sh:prefix "study" ;
sh:namespace "https://w3id.org/phuse/study#"^^xsd:anyURI ;
]
] ;
sh:select
"""SELECT $this (?beginDate AS ?intervalStart) (?endDate AS ?intervalEnd)
WHERE {
$this time:hasBeginning ?beginIRI ;
time:hasEnd ?endIRI .
?beginIRI time:inXSDDate ?beginDate .
?endIRI time:inXSDDate ?endDate .
FILTER (! (?endDate >= ?beginDate ))
}""" ;
] .
The report identifies the interval for Animal Subject 99T1 where End Date precedes Start Date.
a sh:ValidationResult ; sh:sourceConstraint _:bnode_cacffc33_62e3_4c8b_bdba_e71e398a23dc_29 ; sh:sourceShape :SD1002RuleShape ; sh:resultMessage "Interval Start Date on or before End Date. [SD1002]" ; sh:value cj16050:Interval_Animal_184f16eb ] sh:sourceConstraintComponent sh:SPARQLConstraintComponent ; sh:resultSeverity sh:Violation ; sh:focusNode cj16050:Interval_Animal_184f16eb
SPARQL traces the interval back to the AnimalSubject and date values.
SELECT ?animalLabel (?beginDate AS ?intervalStart) (?endDate AS ?intervalEnd)
WHERE {
?animalSubjectIRI study:hasReferenceInterval ?intervalIRI ;
skos:prefLabel ?animalLabel .
?intervalIRI time:hasBeginning ?beginIRI .
?beginIRI time:inXSDDate ?beginDate .
?intervalIRI time:hasEnd ?endIRI .
?endIRI time:inXSDDate ?endDate .
FILTER (! (?endDate >= ?beginDate ))
}
Verify
Verification confirms Animal Subject 99T1 and 99T2 with End Data preceding Start Date. Note how when the start date is a string it also flags AnimalSubject 99T10 as a violator. The SPARQL statement is very similar to the query used in the SHACL-SPARQL constraint.
# RC4 : Verify :
SELECT ?animalLabel (?beginDate AS ?intervalStart) (?endDate AS ?intervalEnd)
WHERE {
?animalSubjectIRI study:hasReferenceInterval ?intervalIRI ;
skos:prefLabel ?animalLabel .
?intervalIRI time:hasBeginning ?beginIRI .
?beginIRI time:inXSDDate ?beginDate .
?intervalIRI time:hasEnd ?endIRI .
?endIRI time:inXSDDate ?endDate .
FILTER (! (?endDate >= ?beginDate ))
}
animalLabel intervaIRI intervalStart intervalEnd "Animal 99T1" cj16050:Interval_Animal_184f16eb 2016-12-07 2016-12-06 "Animal 99T10" cj16050:Interval_Animal_56cbc8c2 "6-DEC-16" 2016-12-07 "Animal 99T2" cj16050:Interval_Animal_21316392 2016-12-08 2016-12-07
Animal Subject Shape - Demographics Domain
Age
The following figure shows the connection from the Animal Subject IRI to its Age value.
Animal Subject Data Structure for Age
The spreadsheet FDA-Validator-Rules.xlsx defines numerous rules associated with Age in the DM domain. This project defines only a subset of these rules as SHACL Shapes. For example, the rule SD2019 “Invalid value for AGETXT” is not applicable because the example study collects AGE (numeric) and not AGETXT (age range as a string).
FDA Rule SD0084
| FDA Validator Rule ID | FDA Validator Message | Business or Conformance Rule Validated | FDA Validator Rule |
|---|---|---|---|
| SD0084 | Negative value for age | Values for age variables cannot be negative, | The value of Age (AGE) cannot be less than 0. |
1. AGE must be greater than or equal to 0.
Data StructureRefer back to previous sections to see how age is indirectly associated with an AnimalSubject via a study:participatesIn predicate that leads to an outcome IRI that in turn contains the age value and units. Most subjects in the study are the same age (8 Weeks), resulting in a small number of tests in outcome IRIs instead of traditional tests on each age value associated with an Animal Subject.
Rule Component 1. AGE must be greater than or equal to 0.age sh:minInclusive 0.
The age for Animal Subject 99T1 was set to -10 for testing.
cj16050:Animal_184f16eb
a study:AnimalSubject ;
study:participatesIn cj16050:AgeDataCollection_Animal_184f16eb,
...
cj16050:AgeDataCollection_Animal_184f16eb
a code:AgeDataCollection ;
code:outcome cj16050:Age_-10_WEEKS .
cj16050:Age_-10_WEEKS
a study:Age ;
time:hasXSDDuration "P56D"^^xsd:duration ;
time:numericDuration -10 ;
time:unitType time:unitWeek .
The shape tests the following condition:
- An Age value must be >= 0
study:hasminInclusive0Shape-Age a sh:NodeShape ;
sh:targetClass study:Age ;
sh:name "agGTE0" ;
sh:description "Age must be greater than or equal to 0." ;
sh:message "Negative value for AGE. [SD0084]" ;
sh:property [
sh:path time:numericDuration ;
sh:minInclusive 0 ;
] .
The report correctly identifies the value ‘-10’.
a sh:ValidationReport ;
sh:result [
a sh:ValidationResult ;
sh:sourceConstraintComponent sh:MinInclusiveConstraintComponent ;
sh:sourceShape [] ;
sh:focusNode cj16050:Age_-10_WEEKS ;
sh:resultPath time:#numericDuration ;
sh:value -10 ;
sh:resultSeverity sh:Violation
] ;
sh:conforms false
The report lists the Age value and Age outcome IRI, but not the AnimalSubject associated with the offending value. SPARQL can be used to identify the Animal Subject using the Age Outcome IRI identified in the report. Source file: /SPARQL/Animal-Age-LT0.rq
SELECT ?animalLabel
WHERE{
?ageDataCollIRI code:outcome cj16050:Age_-10_WEEKS .
?AnimalSubjectIRI study:participatesIn ?ageDataCollIRI ;
skos:prefLabel ?animalLabel .
}
SPARQL independently verifies the Animal Subject with age < 0. Source file: /SPARQL/Animal-Age-LT0.rq
SELECT ?animalIRI ?animalLabel ?age
WHERE{
?animalIRI study:participatesIn ?ageDataCollIRI ;
skos:prefLabel ?animalLabel .
?ageDataCollIRI code:outcome ?ageIRI .
?ageIRI time:numericDuration ?age .
FILTER (?age < 0)
}
Rule SD1121 : Age
| FDA Validator Rule ID | FDA Validator Message | Business or Conformance Rule Validated | FDA Validator Rule |
|---|---|---|---|
| SD1121 | Age or age range must be provided for all subjects, except for Screen Failures. | Age or age range must be provided for all subjects, except for Screen Failures. | Value for Age (AGE) or Age Range (AGETXT) variables should be populated for all subjects with only exception for Screen Failures (ARMCD=SCRNFAIL) and Not Assigned (ARMCD=NOTASSGN) subjects. |
1.AGE value must be present for all subjects that are not screen failures and not assigned to a treatment arm.
Source Data Data StructureThe rule states the age can be missing in cases where the subject is either:
- A screen failure (
armcd=SCRNFL) - Not assigned to a treatment arm (
armcd=NOTASSGN)
Here we encounter another problem in the SDTM data model. SEND data identifies screen failures as armcd='SCRNFL' and subjects not assigned to a treatment arm as armcd='NOTASSGN'. armcd implies it contains values indicating the treatment arm, yet values SCRNFL and NOTASSGN identify subjects who were never assigned to a treatment arm! The data model must change to model these cases without ambiguity.
Screen Failures had an eligibility assessment with an outcome that deemed the subject ineligible for the study (eligibility = FALSE).
Triples would appear similar to:
cj16050:EligibilityDetermination_XXXXXX rdf:type study:EligibilityDetermination ; code:outcome code:RuleOutcome_FALSE . cj16050:Randomization_Animal_2 rdf:type study:Randomization ; skos:prefLabel "Randomization 2" ; code:outcome
Missing Age values are represented in the data has having no object associated with the time:numericduration predicate.
Non missing value of ‘8’:
cj16050:Age_8_WEEKS a study:Age ; skos:prefLabel “Age 8 WEEKS”^^xsd:string ; time:hasXSDDuration “P56D”^^xsd:duration ; time:numericDuration 8 ; time:unitType time:unitWeek .
Missing age value: cj16050:Age_ skos:prefLabel “Age WEEKS”^^xsd:string ; time:hasXSDDuration “P56D”^^xsd:duration ; time:unitType time:unitWeek .
TEMPLATE FOR NEW SECTIONSage ADD HERE 0.
The age for Animal Subjects 99TXX, 99TXX, 99TXX was set to missing.
ADD data for one test case here.
The shape tests the following condition:
- ADD CONDITION
ADD SHACL
The report correctly identifies XXXX.
ADD REPORT EXCERPT
The report ADD DETAILS ABOUT REPORT VALIDATION USING SPARQL. SPARQL can be used to identify Add Source file: Add/SPARQL/foo</font>
ADD
armcd has the value NOTASSGN when there is no randomization outcome. The data conversion process attempts to create the randomization triple as:
The triple is not created due to the missing data for the Object.
The triples for a NOTASSGN subject appear as:
ADD TRIPLES