Graphium Chrysalis: Exploiting Graph Database
0 likes928 views
The document presents a study on Graphium, a tool for analyzing RDF graphs using graph database engines. It details various graph invariants such as reciprocity, density, and h-index, which are crucial for understanding relationships and data quality in graph structures. The findings emphasize the relevance of these invariants in identifying hidden properties and potential associations in semantic data management.
![9
Benchmark of Graph
Graph Name #Nodes #Edges Density #Labels
DSJC1000.1
[Johnson91]
1,000 99,258 0.099 1
DSJC1000.5
[Johnson91]
1,000 499,652 0.50 1
DSJC1000.9
[Johnson91]
1,000 898,898 0.899 1
USA-road-
d.NY
264,346 730,100 0.00001045 7,970
USA-road-
d.FLA
1,070,376 2,687,902 0.00000235 22,704
Berlin10M 2,743,235 9,709,119 0.00000129 40
[Johnson91] Johnson, D., Aragon, C., McGeoch, L., and Schevon, C. Optimization by simulated annealing: an experimental
evaluation; part ii, graph coloring and number partitioning. Operations research 39, 3 (1991), 378–406.
USA-road-d* Graphs 9th DIMACS Implementation Challenge - Shortest Paths http://www.dis.uniroma1.it/challenge9/download.shtml
Berlin10M: Berlin Bechmark-http://wifo5-03.informatik.uni-mannheim.de/bizer/berlinsparqlbenchmark/
COLD 2013](https://image.slidesharecdn.com/graphiumgraph-tamevidal-150323085040-conversion-gate01/85/Graphium-Chrysalis-Exploiting-Graph-Database-9-320.jpg)
Graphium Chrysalis: Exploiting Graph Database
- 1. Graphium Chrysalis: Exploiting Graph Database Engines to Analyze RDF Graphs Alejandro Flores Maria-Esther Vidal Guillermo Palma Universidad Simón Bolívar 1Graph-TA 2015
- 2. Agenda Motivation Graphium Graph Invariants in Graphium Graph-TA 2015 2
- 3. Resource Description Framework (RDF) Model 3 Subject Object Predicate
- 4. Resource Description Framework (RDF) Model 4 duration duration Properties and Relationships are represented as predicates The Beatles Let it be Revolver Help! created 1970 35:16 1965 year 1966 35:01 Liverpool thebeatles.com Subject Object Predicate Source: “Scaling Up Linked Data”. EUCLID project.
- 5. Semantic Data Management RDF Graphs RDF Engines S P O S OP PSO POS OSP OPS SPARQL queries that represent Graph patterns
- 6. Property Graph Model 6 Nodes and edges may have properties Properties: Key-value pairs The Beatles Let it be Revolver Help! created Year: 1970 Duration: 35:16 Year: 1965 Year: 1966 Duration: 35:01 Homepage: thebeatles.com Origin: Liverpool Source: “Scaling Up Linked Data”. EUCLID project.
- 7. Semantic Data Management Property Graphs Graph Database Engines Edges & Nodes Neighborhoods Graph-based tasks
- 8. Semantic Data Management RDF Graphs RDF Engines S P O S OP PSO POS OSP OPS Property Graphs Graph Database Engines SPARQL queries that represent Graph patterns Edges & Nodes Neighborhoods Graph-based tasks
- 9. 9 Benchmark of Graph Graph Name #Nodes #Edges Density #Labels DSJC1000.1 [Johnson91] 1,000 99,258 0.099 1 DSJC1000.5 [Johnson91] 1,000 499,652 0.50 1 DSJC1000.9 [Johnson91] 1,000 898,898 0.899 1 USA-road- d.NY 264,346 730,100 0.00001045 7,970 USA-road- d.FLA 1,070,376 2,687,902 0.00000235 22,704 Berlin10M 2,743,235 9,709,119 0.00000129 40 [Johnson91] Johnson, D., Aragon, C., McGeoch, L., and Schevon, C. Optimization by simulated annealing: an experimental evaluation; part ii, graph coloring and number partitioning. Operations research 39, 3 (1991), 378–406. USA-road-d* Graphs 9th DIMACS Implementation Challenge - Shortest Paths http://www.dis.uniroma1.it/challenge9/download.shtml Berlin10M: Berlin Bechmark-http://wifo5-03.informatik.uni-mannheim.de/bizer/berlinsparqlbenchmark/ COLD 2013
- 10. Adjacency Tests 10 Triple Pattern based Tests
- 11. K-Hop Tests 11
- 12. Semantic Data Management RDF Graphs RDF Engines S P O S OP PSO POS OSP OPS Property Graphs Graph Database Engines SPARQL queries that represent Graph patterns Edges & Nodes Neighborhoods Graph-based tasks
- 13. 13 GRAPHIUM Neo4j Sparksee Graph-based API RDF-based API GRAPHIUM: http://graphium.ldc.usb.ve
- 15. 15 GRAPHIUM Neo4j Sparksee Graph-based API RDF-based API Data Mining Traversal API Graph Invariants GRAPHIUM: http://graphium.ldc.usb.ve
- 16. 16
- 17. Graph Invariants 17 Invariant Description Vertex and Edge Count number of vertices and edges in the graph. Graph Density number of edges in the graph divided by the number of possible edges in a complete digraph. Reciprocity Reciprocity measures the extend to which a triple that relates resources A and B is reciprocated by a another triple that relates B with A too. In- and Out-degree Distribution Distribution of the number of in-coming and out-going edges of the vertices of a graph. In-coming and Out-going H-index h is the maximum number, such that h vertices have each at least h in-coming neighbors (resp., out-going neighbors) in the graph.
- 18. Reciprocity: Reciprocal edges indicates stronger relationships between vertices. Graph invariants diseasome:possibleDrug Drugbank Diseasome drugbank:possibleDiseaseTarget
- 19. Reciprocity: Reciprocal edges indicates stronger relationships between vertices. Graph invariants Drugbank Diseasome diseasome:possibleDrug drugbank:possibleDiseaseTarget drugbank:DB00157 drugbank:possibleDiseaseTarget diseasome:diseases/0 diseasome:diseases/1 diseasome:diseases/4198 … diseasome:diseases/0 diseasome:possibleDrug drugbank:DB00157 diseasome:diseases/1 drugbank:DB00157 Reciprocity values less than 1.0 indicates that there are drugs associated with diseases that do not have their reciprocal link.
- 20. Reciprocity: Reciprocal edges indicates stronger relationships between vertices. Graph invariants Drugbank Diseasome diseasome:possibleDrug drugbank:possibleDiseaseTarget drugbank:DB00157 drugbank:possibleDiseaseTarget diseasome:diseases/0 diseasome:diseases/1 diseasome:diseases/4198 … diseasome:diseases/0 diseasome:possibleDrug drugbank:DB00157 diseasome:diseases/1 drugbank:DB00157 Reciprocity can be used to determine Data Quality and Completeness
- 21. H-Index Sets
- 22. H-Index Set Out S1 O1 P1 S2 O2 P2 S3 O3 P3 P4 P5 S4 P6 O4 P7 P8 A set F of vertices, where H is the maximum number, such that the vertices in F have each at least H out-going neighbors. S5
- 23. H-Index Set Out S1 O1 P1 S2 O2 P2 S3 O3 P3 P4 P5 S4 P6 O4 P7 P8 A set F of vertices, where 2 is the maximum number, such that the vertices in F have each at least 2 out-going neighbors. S5 F={S1,S2,S3} 3 3 2
- 24. H-Index Set In S1 O1 P1 S2 O2 P2 S3 O3 P3 P4 P5 S4 P6 O4 P7 P8 A set F of vertices, where H is the maximum number, such that the vertices in F have each at least H in-coming neighbors. S5
- 25. H-Index Set Out S1 O1 P1 S2 O2 P2 S3 O3 P3 P4 P5 S4 P6 O4 P7 P8 A set F of vertices, where 3 is the maximum number, such that the vertices in F have each at least 3 in-coming neighbors. S5 F={O1,O2,O3} 3 3 3
- 26. Graph invariants SELECT DISTINCT * WHERE { ?s drugbank:drugCategory <http://wifo5-04.informatik.uni-mannheim.de/drugbank/resource/drugcategory/micronutrient>. ?s drugbank:target ?o. ?o drugbank:drugReference ?o2. ?o drugbank:goClassificationComponent ?o3 } Drugbank SPARQL endpoint times out “References and GO annotations of the targets associated with the Micro Nutrient Drugs”
- 27. 48 Drugs 685 Targets … … … … … … … … … … References GO Terms 2-hop of Micro Nutrient Drugs Graph invariants
- 28. 48 Drugs 685 Targets … … … … … … … … … … References GO Terms Graph invariants
- 29. 48 Drugs 685 Targets … … … … … … … … … … References GO Terms 10 Drugs have at least 57 out-going links H-Index Out
- 30. 48 Drugs 685 Targets … … … … … … … … … … References GO Terms 47 Targets have at least 57 out-going links H-Index Out
- 31. 48 Drugs 685 Targets … … … … … … … … … … References GO Terms 6 References have at least 21 in-coming links H-Index In H-Index Sets can be used to explain query complexity
- 32. H-Index Sets to Validate Potential Novel Associations
- 33. H-Index Sets Network of Targets and Drugs Targets Drugs
- 34. H-Index Sets 34 A set F of targets, where H is the maximum number, such that the targets in F have each at least H out-going neighbors. Targets
- 35. H-Index Sets 35 A set F of targets, where H is the maximum number, such that the targets in F have each at least H out-going neighbors. A set F of drugs, where H is the maximum number, such that the drugs in F have each at least H in- coming neighbors. Targets Drugs
- 36. Set of Targets and Drugs 900 Drugs, 1,000 Targets and 5,000 Interactions: Nuclear receptor, Gprotein- coupled receptors (GPCRs), Ion channels, and Enzymes. DrugBank K. Bleakley and Y. Yamanishi. Supervised prediction of drug target interactions using bipartite local models. Bioinformatics, 25(18).2009. 36 GPCR Drugs 223 Targets 95 Interactions 635 Avg Interaction per Target 6.68 Avg Interaction per Drug 2.84
- 37. Drugbank Drugs in the dataset of Gprotein-coupled receptors (GPCRs) H-index Out is 14 15 Targets are in the H-Index Set Out F={hsa:1128, hsa:1129, hsa:146, hsa:147, hsa:148, hsa:150, hsa:151,hsa:152,hsa:153,hsa:154,hsa:155,hsa:1812, hsa:1813, has:3269,has:3356}
- 38. Drugbank Drugs in the dataset of Gprotein-coupled receptors (GPCRs) H-index Out is 14 15 Targets are in the H-Index Set Out
- 39. H-Index Sets
- 40. D02076 hsa:146 D02076 hsa:147 D00604 has:147 Belong to the H-index Set Associations between Drugs and Targets that are not in Drugbank Validated in STICTH http://stitch.embl.de/ H-Index Sets
- 41. D02076 hsa:146 D02076 hsa:147 D00604 has:147 Belong to the H-index Set Associations between Drugs and Targets that are not in Drugbank Validated in STICTH http://stitch.embl.de/ H-Index Sets can be used to Validate the Discovered Associations H-Index Sets
- 43. Conclusions Graph Invariants: Remain the same under two isomorphic graphs and any representation. Allow for uncovering hidden properties of the graphs Reciprocity Density H-Index Set Reciprocity can suggest data quality and incompleteness. Density can be used to explain complexity of graph tasks H-index sets can comprise entities useful to discover potential novel associations.