Docstoc

uml definition

Document Sample
uml definition Powered By Docstoc
					              VCDE Silver Level Compatibility Review
                 GenePattern – Detailed Report

                                          Jan 22, 2008

(Report format)
UML Class Name                                    CDE Object Name
    class comments here

UML Attribute Name                                CDE Property Name
   attribute comments here



ClassMembership                             Member Class
 The UML definition is more detailed than the concepts. I suggest adding a concept for mapping to
   improve the semantics. The attribute should have been named ClassMembershipMapping.

classNames                                    Class Name
 Inconsistent datatypes. Java primitive type “char” used in model though corresponding CDE has
    “java.lang.String” datatype. Primitive datatypes are not recommended datatypes.

numSamples                                   Specimen Number
 Semantics ok.
 Inconsistent datatypes. Java primitive type “int” used in model though corresponding CDE has
   “java.lang.Integer” datatype. Primitive datatypes are not recommended datatypes.


sampleClasses                                Specimen Class
 Ok although I would have liked to have seen an explicit concept for index (it was implied).
 Inconsistent datatypes. Java primitive type “int” used in model though corresponding CDE has
   “java.lang. Integer” datatype. Primitive datatypes are not recommended datatypes.


sampleNames                                   Specimen Name
Inconsistent datatypes. Java primitive type “char” used in model though corresponding CDE has
“java.lang.String” datatype. Primitive datatypes are not recommended datatypes

Cluster                                    Cluster
 No issue to report.

itemName                                   Entity Name
 No issue to report.
name                                         Name
 No issue to report.

ComparativeMarkerSelectionResultCollection          Comparative Marker Selection Outcome
                                                    Container Object
   OK


ComparativeMarkerSelectionParameterSet              Comparative Marker Selection Parameter Set


balanced                                     Balanced Permutation

   Attribute name does not reflect definition. Suggested attribute name “balancedPermutation”.

complete                                     Complete Flag

   Annotated using concepts-flag: C43578 and complete: C25250. Used annotations do not reflect
    the meaning assigned in the attribute definition “whether to perform all possible permutations”.
    There is no mention of “permutation” in the concept used for annotations.

minStd                                       Minimum Standard Deviation

   Abbreviation used in attribute name. Unabbreviated name usage suggested.

numberOfPermutations                         Number Permutation

   Attribute name is plural. Singular name usage suggested.

   Inconsistent datatypes. Java primitive type “int” used in model though corresponding CDE has
    “java.lang.Integer” datatype. Primitive datatypes are not recommended datatypes.

phenotypeTest                                Phenotype Test

   Attribute name type in the UML model is “char” but the value domain datatype provided is
    “java.lang.String”. Please see SIW 3.1 report below-




randomSeed                                   Random Seed

   Inconsistent datatypes. Java primitive type “int” used in model though corresponding CDE has
    “java.lang.Integer” datatype. Primitive datatypes are not recommended datatypes.

smoothPvalues                                P-Value Smoothing
   Attribute name is plural. Singular name usage suggested.


testDirection                               Test Direction

   Annotated using concepts-direction: C42677 and test: C47891. Used annotations do not reflect the
    meaning assigned in the attribute definition “the test to perform (upregulated for class0, up
    regulated for class 1, two sided)”.

   Attribute name type in the UML model is “char” but the value domain datatype provided is
    “java.lang.String”. Please see SIW 3.1 report below-




testStatistic                               Statistical Test

   Attribute name type in the UML model is “char” but the value domain datatype provided is
    “java.lang.String”. Please see SIW 3.1 report below-




ConsensusClusteringParameterSet             Consensus Clustering Parameter Set
 No issues to report

algorithm                              Algorithm
 Value domain name “ConsensusClusteringAlgorithms” do not conform with value domain naming
    conventions.

descentIterations                            Self-Organizing Map Iteration
 Attribute name is plural. Singular name usage is suggested.
 Inconsistent datatypes. Java primitive type “int” used in model though corresponding CDE has
   “java.lang.Integer” datatype. Primitive datatypes are not recommended datatypes.

kmax                                         Cluster Maximum Number
 Inconsistent datatypes. Java primitive type “int” used in model though corresponding CDE has
   “java.lang.Integer” datatype. Primitive datatypes are not recommended datatypes.

mergeType                              Merge Type
 Value domain name “ConsensusClusteringMergeType” do not conform with value domain naming
   conventions.

niter                                       Iteration Number
   Annotated using concept “C25337:Number:A numeral or string of numerals expressing value,
    quantity, or identification”. Concept “C25463:Count: To determine the number or amount of
    something; the result of this activity.” seems to be a better fit for annotation.
   Inconsistent datatypes. Java primitive type “int” used in model though corresponding CDE has
    “java.lang.Integer” datatype. Primitive datatypes are not recommended datatypes.

normIterations                               Normalize Iteration
 Concept “C25463:Count: To determine the number or amount of something; the result of this
   activity.” missing from definition. The definition “number of row/column normalization iterations”
   clearly suggests it is a count/number.
 Attribute name is plural. Singular name usage is suggested.
 Inconsistent datatypes. Java primitive type “int” used in model though corresponding CDE has
   “java.lang.Integer” datatype. Primitive datatypes are not recommended datatypes.


normalizeType                                 Normalize Type
 No issues to report

resample                                    Sampling Repeat
 Based on the definition “resampling scheme ('subsample[ratio]', 'features[nfeat]', 'nosample')”, it
    should have a enumerated value domain.
 Base on the definition “resampling scheme ('subsample[ratio]', 'features[nfeat]', 'nosample')”, rather
    than “C47886: Repeat” the primary concept should be “C25665:Scheme: An elaborate and
    systematic plan of action.” or “C16847:Technique: A practiced and regimented skill or series of
    actions.”

ConsensusClusterResultCollection              Consensus Clustering Outcome Container Object
 No issues to report

ConsensusMatrix                               Consensus Matrix
 No issues to report

ConsensusMatrixRow                            Consensus Matrix Row

   Annotated using concept “C45805:Matrix: Ground substance in which things are embedded or that
    fills a space (e.g., extracellular matrix or nuclear matrix).”. Concept “C45802: A rectangular array
    of elements (or entries) set out by rows and columns.” seems to be a better fit for annotation.

feature                                 Characteristics
 Probably concept “C44283:Array_Feature:A specific position on a microarray.” can be used to
    provide a more specific annotation.

scores                                       Score
 Attribute name is plural. Singular name usage is suggested.
 Inconsistent datatypes. Java primitive type “double” used in model though corresponding CDE has
    “java.lang.Double” datatype. Primitive datatypes are not recommended datatypes.
MarkerResult                              Biomarker Outcome
 “Biomarker Outcome” does not precisely denote the concept of this UML class. Does EVS not
  have words such as “marker” and “result”?


Bonferroni                                Bonferroni Adjustment
 OK

FDR                                       False Discovery Rate
 OK

FWER                                      Family-Wise Error Rate
 OK

QValue                                    Q-Value
 OK

class0Mean                                First State Mean
 UML gives no definition, assuming this is a well known concept. CDE needs to be more precise
    about “first state”

class0Std                                 First State Standard Deviation
 Same as the above

class1Mean                                Second State Mean
 Same as the above

class1Std                                 Second State Standard Deviation
 Same as the above

description                               Description
 OK

feature                                   Characteristics
 CDE should just use “feature” as the name.

featureP                                  Characteristic P-Value
 OK

featurePHigh                              Maximum Characteristic P-Value
 OK

featurePLow                               Minimum Characteristic P-Value
 OK

foldChange                                Fold Change
 UML needs a better definition
K                                             Number
 CDE too generic

maxT                                          MaxT
 OK

rank                                          Rank
 OK

score                                         Score
 OK

PreprocessDatasetParameterSet              Data Processing Data Set Parameter Set
 Ok, but I would have preferred to see explicit concept(s) for analysis service (it is implied).

ceiling                                       Default Most Value
 Semantics ok.
 Inconsistent datatypes. Java primitive type “float” used in model though corresponding CDE has
    “java.lang.Float” datatype. Primitive datatypes are not recommended datatypes.

columnThreshold                               Column Threshold
 Semantics ok.
 Inconsistent datatypes. Java primitive type “float” used in model though corresponding CDE has
    “java.lang.Float” datatype. Primitive datatypes are not recommended datatypes.

filterFlag                                    Filter Flag
 It would have been nice to see concepts for the type of flag and filter such as, variation and
     thresholding.
 Inconsistent datatypes. Java primitive type “boolean” used in model though corresponding CDE
     has “java.lang.Boolean” datatype. Primitive datatypes are not recommended datatypes.

logBaseTwo                                  Log2
 Semantics ok.
 Inconsistent datatypes. Java primitive type “boolean” used in model though corresponding CDE
   has “java.lang.Boolean” datatype. Primitive datatypes are not recommended datatypes.

maxSigmaBinning                              Bin Standard Deviation Maximum
 Semantics ok.
 Inconsistent datatypes. Java primitive type “int” used in model though corresponding CDE has
   “java.lang.Integer” datatype. Primitive datatypes are not recommended datatypes.

minChange                                    Minimum Fold Change
 I would have liked to have seen a concept for filter.
 Inconsistent datatypes. Java primitive type “float” used in model though corresponding CDE has
   “java.lang.Float” datatype. Primitive datatypes are not recommended datatypes.

minDelta                                      Minimum Delta
   I would have liked to have seen a concept for filter.
   Inconsistent datatypes. Java primitive type “float” used in model though corresponding CDE has
    “java.lang.Float” datatype. Primitive datatypes are not recommended datatypes.

numExclude                                   Exclusion Number
 Detailed UML definition while concepts are general.
 Inconsistent datatypes. Java primitive type “int” used in model though corresponding CDE has
   “java.lang.Integer” datatype. Primitive datatypes are not recommended datatypes.

numberOfColumnsAboveThreshold                Greater Than Threshold Column Number
 Concepts do not quite capture the UML definition in regards to removing a row if conditions are
   met.
 Inconsistent datatypes. Java primitive type “int” used in model though corresponding CDE has
   “java.lang.Integer” datatype. Primitive datatypes are not recommended datatypes.

preprocessingFlag                            Data Processing Flag
 I would have liked to see more specific concepts for Discretization and normalization.
 Enumerations are good!
 Inconsistent datatypes. Java primitive type “char” used in model though corresponding CDE has
   “java.lang.String” datatype. Primitive datatypes are not recommended datatypes.
 Value domain name “PreprocessDatabasePreprocessingFlag” does not conform with the value
   domain naming conventions.

probabilityThreshold                         Probability Threshold
 Would have like to have seen a concept for filter.
 Inconsistent datatypes. Java primitive type “float” used in model though corresponding CDE has
   “java.lang.Float” datatype. Primitive datatypes are not recommended datatypes.

threshold                                     Threshold Less Than Value
 Inconsistent datatypes. Java primitive type “float” used in model though corresponding CDE has
    “java.lang.Float” datatype. Primitive datatypes are not recommended datatypes.