Skip to content

Latest commit

 

History

History
354 lines (323 loc) · 9.75 KB

qualifier_example_data_and_queries.md

File metadata and controls

354 lines (323 loc) · 9.75 KB
Raw

Data (Source Record Statements)

Plain language expression of what each record asserts to be true

  • Record A: Fenofibrate binds to PPARA protein (CTD)
  • Record B: Fenofibrate results in increased activity of PPARA protein (CTD)
  • Record C: Fenofibrate results in increased expression of PPARA (CTD)
  • Record D: Bezofuran increases the activity of PPARA (CTD)
  • Record E: Bezofuran results in decreased degradation of PPARA (CTD)
  • Record F: Okadaic Acid results in increased phosphorylation of PPARA (CTD)

TO DO: Add examples to support a drug prediction reasoning path: Chemical - negatively regulates (dec activity of) - Gene/Protein - physically interacts with - Gene/Protein - causes/gene assoc with - Condition

  • Mifepristone results in decreased activity of PPARA protein
  • PPARA protein physically interacts with GRB2
  • GRB2 associated with Wiskott-Aldrich Syndrome

Structured Bioink Representations

Record A: Fenofibrate binds to PPARA protein (CTD)

subject: Fenofibrate
predicate: physically interacts with
object: PPARA

Record B: Fenofibrate results in increased activity of PPARA protein (CTD)

subject: Fenofibrate
predicate: affects
qualified_predicate: causes
object: PPARA 
object_aspect: activity
object_direction: increased

Record C: Fenofibrate results in increased expression of PPARA (CTD)

subject: Fenofibrate
predicate: affects
qualified_predicate: causes
object: PPARA 
object_aspect: expression
object_direction: increased

Record D: Bezofuran increases the activity of PPARA (CTD)

subject: Bezofuran 
predicate: affects
qualified_predicate: causes
object: PPARA 
object_aspect: activity
object_direction: increased

Record E: Bezofuran results in decreased degradation of PPARA (CTD)

subject: Bezofuran
predicate: affects
qualified_predicate: causes
object: PPARA
object_aspect: degradation
object_direction: decreased

Record F: Okadaic Acid results in increased phosphorylation of PPARA (CTD)

subject: Okadaic Acid
predicate: affects
qualified_predicate: causes
object: PPARA
object_aspect: phosphorylation
object_direction: increased

TRAPI Queries

Q1: What Chemical Entities affect PPARA?

Expected Results: Fenofibrate, Bezofuran, and Okadoaic Acid - based on records B, C, D, E, F

Structured Query:

{
  "message": {
    "query_graph": {
      "nodes": {
        "n0": {
          "category": "biolink:ChemicalSubstance"
        },
        "n1": {
          "id": "HGNC:9232"
        }
      },
      "edges": {
        "e01": {
          "subject": "n0",
          "object": "n1",
          "predicate": [
            "affects"
          ]
        }
      }
    }
  }
}

Q2: What Chemical Entities increase the activity of PPARA?

Expected Results: Fenofibrate and Bezofuran - based on records B, D

Structured Query:


{
  "message": {
    "query_graph": {
      "nodes": {
        "n0": {
          "category": "biolink:ChemicalSubstance"
        },
        "n1": {
          "id": "HGNC:9232"
        }
      },
      "edges": {
        "e01": {
          "subject": "n0",
          "object": "n1",
          "predicate": [
            "affects"
          ],
          "qualifier_constraints": [
            {
              "qualifier_set": [
                {
                  "qualifier_type_id": "biolink:object_aspect",
                  "value": "activity"
                },
                {
                  "qualifier_type_id": "biolink:object_direction",
                  "value": "increased"
                },
                {
                  "qualifier_type_id": "biolink:qualified_predicate",
                  "value": "biolink:causes"
                }
              ]
            }
          ]
        }
      }
    }
  }
}

Q3: What compounds increase the amount or activity of PPARA?

Expected Results: Fenofibrate and Bezofuran - based on records B, C, D, E

Structured Query:

{
  "message": {
    "query_graph": {
      "nodes": {
        "n0": {
          "category": "biolink:ChemicalSubstance"
        },
        "n1": {
          "id": "HGNC:9232"
        }
      },
      "edges": {
        "e01": {
          "subject": "n0",
          "object": "n1",
          "predicate": [
            "affects"
          ],
          "qualifier_constraints": [
            {
              "qualifier_set": [
                {
                  "qualifier_type_id": "biolink:object_aspect",
                  "value": "activity"
                },
                {
                  "qualifier_type_id": "biolink:object_direction",
                  "value": "increased"
                },
                {
                  "qualifier_type_id": "biolink:qualified_predicate",
                  "value": "biolink:causes"
                }
              ]
            },
            {
              "qualifier_set": [
                {
                  "qualifier_type_id": "biolink:object_aspect",
                  "value": "abundance"
                },
                {
                  "qualifier_type_id": "biolink:object_direction",
                  "value": "increased"
                },
                {
                  "qualifier_type_id": "biolink:qualified_predicate",
                  "value": "biolink:causes"
                }
              ]
            },
            {
              "qualifier_set": [
                {
                  "qualifier_type_id": "biolink:object_aspect",
                  "value": "degradation"
                },
                {
                  "qualifier_type_id": "biolink:object_direction",
                  "value": "decreased"
                },
                {
                  "qualifier_type_id": "biolink:qualified_predicate",
                  "value": "biolink:causes"
                }
              ]
            }
          ]
        }
      }
    }
  }
}

Questions:

  • Is there a way we could infer decreased degradation --> increased abundance so Record E returned even if not explicitly specified? This was easier to do when aspect and direction were composed in a single predicate - by placing 'decreases degradation of' under 'increases abundance of' in the predicate hierarchy. But with aspect and directions now split into separate qualifiers, how might we derive the same inference? Will need more complex rules.

Q4: What compounds increase the activity or phosphorylation of PPARA but not its abundance

Based on Vlado's example query, but modified 'synthesis' -> 'phosphorylation' (as 'synthesis no longer in CTD data), and modified 'expression' -> 'abundance' (to illustrate need for hierarchical inference)

Expected Results: Okadaic Acid (based on record F), but not Fenofibrate or Bezofuran despite records Band E asserting these to increase the activity of PPARA - because Fenofibrate and Bezofuran also increases abundance of PPARA (Fenofibrate via increasing its expression, and Bezofuran via decreasing its degradation)

Structured Query:

{
  "message": {
    "query_graph": {
      "nodes": {
        "n0": {
          "category": "biolink:ChemicalSubstance"
        },
        "n1": {
          "id": "HGNC:9232"
        }
      },
      "edges": {
        "e01": {
          "subject": "n0",
          "object": "n1",
          "predicate": [
            "affects"
          ],
          "qualifier_constraints": [
            {
              "qualifier_set": [
                {
                  "qualifier_type_id": "biolink:object_aspect",
                  "value": "activity"
                },
                {
                  "qualifier_type_id": "biolink:object_direction",
                  "value": "increased"
                },
                {
                  "qualifier_type_id": "biolink:qualified_predicate",
                  "value": "biolink:causes"
                }
              ]
            },
            {
              "qualifier_set": [
                {
                  "qualifier_type_id": "biolink:object_aspect",
                  "value": "synthesis"
                },
                {
                  "qualifier_type_id": "biolink:object_direction",
                  "value": "increased"
                },
                {
                  "qualifier_type_id": "biolink:qualified_predicate",
                  "value": "biolink:causes"
                }
              ]
            },
            {
              "qualifier_set": [
                {
                  "not": "TRUE",
                  "qualifier_type_id": "biolink:object_aspect",
                  "value": "abundance"
                },
                {
                  "qualifier_type_id": "biolink:object_direction",
                  "value": "increased"
                },
                {
                  "qualifier_type_id": "biolink:qualified_predicate",
                  "value": "biolink:causes"
                }
              ]
            },
            {
              "qualifier_set": [
                {
                  "not": "TRUE",
                  "qualifier_type_id": "biolink:object_aspect",
                  "value": "degradation"
                },
                {
                  "qualifier_type_id": "biolink:object_direction",
                  "value": "decreased"
                },
                {
                  "qualifier_type_id": "biolink:qualified_predicate",
                  "value": "biolink:causes"
                }
              ]
            }
          ]
        }
      }
    }
  }
}

Questions:

  • Consider direction of hierarchical inference/query expansion here, when not = TRUE. When not = FALSE, expand queried term down its sub-hierarchies. When not = TRUE, I think we also expand queried term down its sub-hierarchies.