schema.gql

define

##########################
# Entities
##########################
gene sub entity,
    has gene-symbol,
    has entrez-id,
    has ensembl-id,
    has gene-name,
    plays encoding-gene,
    plays analysed-gene,
    plays associated-gene,
    plays target-gene,
    plays differentially-expressed-gene;
protein sub entity,
    has uniprot-id,
    has uniprot-name,
    has protein-name,
    plays encoded-protein,
    plays expressed-protein,
    plays associated-protein,
    plays interacting-protein,
    plays participating-protein,
    plays target-protein,
    plays sequence-similar-protein;
disease sub entity,
    has disease-name,
    has disease-id,
    plays associated-disease,
    plays affected-disease;
drug sub entity,
    has drug-name,
    has drug-chembl-id
    plays inhibitor,
    plays antagonist,
    plays blocker,
    plays therapeutic,
    plays interacted-drug
    plays associated-drug;
tissue sub entity,
    has tissue-name,
    plays enhanced-tissue,
    plays tissue-context,
    plays associated-tissue;
# This relationship can be reused anywhere where we want to record where data was ingested from
database sub entity,
    has database-name,
    plays ingested-source;
geo-series sub entity,
    has GEOStudy-id,
    plays containing-study;
pathway sub entity,
    has pathway-name,
    has pathway-id,
    plays participated-pathway;
geo-comparison sub entity,
	has GEOComparison-id,
	plays compared-groups,
	plays conducted-analysis;

##########################
# Relationships
##########################
comparison sub relationship,
    relates containing-study,
    relates compared-groups;
genetic-analysis sub relationship,
	relates conducted-analysis,
	relates analysed-gene,
	has p-value;
differentially-expressed-finding sub relationship,
	relates conducted-analysis,
    relates differentially-expressed-gene;
protein-sequence-similarity sub relationship,
    relates sequence-similar-protein;
gene-protein-encoding sub relationship,
	relates encoded-protein,
	relates encoding-gene;
protein-disease-association sub relationship,
    relates associated-protein,
    relates associated-disease,
    plays ingested-data,
    has score;
gene-disease-association sub relationship,
    relates associated-gene,
    relates associated-disease,
    plays ingested-data,
    has score;
gene-tissue-association sub relationship,
	relates associated-tissue,
	relates associated-gene;
data-ingestion sub relationship,
    relates ingested-data,
    relates ingested-source;
drug-protein-interaction sub relationship,
    relates inhibitor,
    relates antagonist,
    relates blocker,
    relates target-protein;
drug-gene-interaction sub relationship,
    relates target-gene,
    relates interacted-drug;
tissue-enhancement sub relationship,
    relates enhanced-tissue,
    relates expressed-protein;
process-localisation sub relationship,
    relates biomolecular-process,
    relates tissue-context;
molecular-interaction sub relationship,
    relates interacting-molecule;
protein-protein-interaction sub molecular-interaction,
    relates interacting-protein as interacting-molecule,
    plays biomolecular-process;
drug-disease-association sub relationship,
    relates affected-disease,
    relates therapeutic;
pathway-participation sub relationship,
	relates participating-protein,
	relates participated-pathway;
disease-tissue-association sub relationship,
    relates associated-disease,
    relates associated-tissue;
drug-tissue-association sub relationship,
    relates associated-drug,
    relates associated-tissue;

##########################
# Attributes
##########################
name sub attribute datatype string;
disease-name sub name;
database-name sub name;
uniprot-name sub name;
gene-symbol sub name;
gene-name sub name;
protein-name sub name;
pathway-name sub name;
tissue-name sub name;
drug-name sub name;

identifier sub attribute datatype string;
disease-id sub identifier;
protein-id sub identifier;
uniprot-id sub protein-id;
entrez-id sub protein-id;
ensembl-id sub protein-id;
GEOStudy-id sub identifier;
GEOComparison-id sub identifier;
pathway-id sub identifier;
drug-chembl-id sub identifier;

p-value sub attribute datatype double;
score sub attribute datatype double;

##########################
# Rules
##########################
when-p-value-sufficient-then-gene-differentially-expressed sub rule,
when {
	$gc isa geo-comparison;
	$g isa gene;
	$gn (conducted-analysis: $gc, analysed-gene: $g) isa genetic-analysis, has p-value $v;
    $v <= 0.1;
} then {
    (differentially-expressed-gene: $g, conducted-analysis: $gc) isa differentially-expressed-finding;
};

gene-disease-association-and-gene-protein-encoding-protein-disease-association sub rule,
when {
    $g isa gene;
    $pr isa protein;
    $di isa disease;
    (associated-disease: $di, associated-gene: $g) isa gene-disease-association;
    (encoding-gene: $g, encoded-protein: $pr) isa gene-protein-encoding;
} then {
    (associated-protein: $pr, associated-disease: $di) isa protein-disease-association;
};

gene-tissue-association-gene-protein-encoding-implies-tissue-enhancement sub rule,
when {
    $g isa gene;
    $pr isa protein;
    $ti isa tissue;
    (associated-tissue: $ti, associated-gene: $g) isa gene-tissue-association;
    (encoding-gene: $g, encoded-protein: $pr) isa gene-protein-encoding;
} then {
    (expressed-protein: $pr, enhanced-tissue: $ti) isa tissue-enhancement;
};

drug-gene-interaction-and-gene-protein-encoding-implies-drug-protein-interaction sub rule,
when {
    $g isa gene;
    $pr isa protein;
    $dr isa drug;
    ($dr, target-gene: $g) isa drug-gene-interaction;
    (encoding-gene: $g, encoded-protein: $pr) isa gene-protein-encoding;
} then {
    (target-protein: $pr, inhibitor: $dr) isa drug-protein-interaction;
};

protein-disease-association-and-protein-sequence-similarity-implies-drug-disease-association sub rule,
when {
	$di isa disease;
	$pr isa protein;
	$pr2 isa protein;
	$pr != $pr2;
	$dr isa drug;
	(associated-disease: $di, associated-protein: $pr) isa protein-disease-association;
	(sequence-similar-protein: $pr, sequence-similar-protein: $pr2) isa protein-sequence-similarity;
	(target-protein: $pr2, $dr) isa drug-protein-interaction;
} then {
	(affected-disease: $di, therapeutic: $dr) isa drug-disease-association;
};

protein-disease-association-and-tissue-enhancement-implies-disease-tissue-association sub rule,
when {
   $di isa disease;
   $pr isa protein;
   $ti isa tissue;
   (associated-disease: $di, associated-protein: $pr) isa protein-disease-association;
   (expressed-protein: $pr, enhanced-tissue: $ti) isa tissue-enhancement;
} then {
   (associated-disease: $di, associated-tissue: $ti) isa disease-tissue-association;
};

drug-gene-interaction-and-gene-tissue-association-implies-drug-tissue-association sub rule,
when {
   $dr isa drug;
   $ti isa tissue;
   $ge isa gene;
   (interacted-drug: $dr, target-gene: $ge) isa drug-gene-interaction;
   (associated-gene: $ge, associated-tissue: $ti) isa gene-tissue-association;
} then {
   (associated-drug: $dr, associated-tissue: $ti) isa drug-tissue-association;
};