Merge pull request #195 from prasham21/browse-statements

Add ability to browse statements underlying gene set analysis results

src/indra_cogex/apps/gla/gene_blueprint.py

@@ -164,6 +164,8 @@ def discretize_analysis():
             errors=errors,
             method=form.correction.data,
             alpha=form.alpha.data,
+            minimum_evidence=form.minimum_evidence.data,
+            minimum_belief=form.minimum_belief.data,
             go_results=results["go"],
             wikipathways_results=results["wikipathways"],
             reactome_results=results["reactome"],

src/indra_cogex/apps/queries_web/__init__.py

@@ -83,22 +83,29 @@
 gene_expression_ns = Namespace("Gene Expression Queries", CATEGORY_DESCRIPTIONS['gene_expression'], path="/api")
 go_terms_ns = Namespace("GO Terms Queries", CATEGORY_DESCRIPTIONS['go_terms'], path="/api")
 clinical_trials_ns = Namespace("Clinical Trials Queries", CATEGORY_DESCRIPTIONS['clinical_trials'], path="/api")
-biological_pathways_ns = Namespace("Biological Pathways Queries", CATEGORY_DESCRIPTIONS['biological_pathways'], path="/api")
+biological_pathways_ns = Namespace("Biological Pathways Queries", CATEGORY_DESCRIPTIONS['biological_pathways'],
+                                   path="/api")
 drug_side_effects_ns = Namespace("Drug Side Effects Queries", CATEGORY_DESCRIPTIONS['drug_side_effects'], path="/api")
 ontology_ns = Namespace("Ontology Queries", CATEGORY_DESCRIPTIONS['ontology'], path="/api")
-literature_metadata_ns = Namespace("Literature Metadata Queries", CATEGORY_DESCRIPTIONS['literature_metadata'], path="/api")
+literature_metadata_ns = Namespace("Literature Metadata Queries", CATEGORY_DESCRIPTIONS['literature_metadata'],
+                                   path="/api")
 statements_ns = Namespace("Statements Queries", CATEGORY_DESCRIPTIONS['statements'], path="/api")
 drug_targets_ns = Namespace("Drug Targets Queries", CATEGORY_DESCRIPTIONS['drug_targets'], path="/api")
 cell_markers_ns = Namespace("Cell Markers Queries", CATEGORY_DESCRIPTIONS['cell_markers'], path="/api")
-disease_phenotypes_ns = Namespace("Disease-Phenotypes Association Queries", CATEGORY_DESCRIPTIONS['disease_phenotypes'], path="/api")
-gene_disease_variant_ns = Namespace("Gene-Disease-Variant Association Queries", CATEGORY_DESCRIPTIONS['gene_disease_variant'], path="/api")
-research_project_output_ns = Namespace("Project-Research Queries", CATEGORY_DESCRIPTIONS['project_research_outputs'], path="/api")
+disease_phenotypes_ns = Namespace("Disease-Phenotypes Association Queries", CATEGORY_DESCRIPTIONS['disease_phenotypes'],
+                                  path="/api")
+gene_disease_variant_ns = Namespace("Gene-Disease-Variant Association Queries",
+                                    CATEGORY_DESCRIPTIONS['gene_disease_variant'], path="/api")
+research_project_output_ns = Namespace("Project-Research Queries", CATEGORY_DESCRIPTIONS['project_research_outputs'],
+                                       path="/api")
 gene_domains_ns = Namespace("Gene Domain Queries", CATEGORY_DESCRIPTIONS['gene_domains'], path="/api")
-phenotype_variant_ns = Namespace("Phenotype-Variant Association Queries", CATEGORY_DESCRIPTIONS['phenotype_variant'], path="/api")
+phenotype_variant_ns = Namespace("Phenotype-Variant Association Queries", CATEGORY_DESCRIPTIONS['phenotype_variant'],
+                                 path="/api")
 drug_indications_ns = Namespace("Drug Indication Queries", CATEGORY_DESCRIPTIONS['drug_indications'], path="/api")
 gene_codependence_ns = Namespace("Gene Codependence Queries", CATEGORY_DESCRIPTIONS['gene_codependence'], path="/api")
 enzyme_activity_ns = Namespace("Enzyme Activity Queries", CATEGORY_DESCRIPTIONS['enzyme_activity'], path="/api")
-cell_line_properties_ns = Namespace("Cell Line Property Queries", CATEGORY_DESCRIPTIONS['cell_line_properties'], path="/api")
+cell_line_properties_ns = Namespace("Cell Line Property Queries", CATEGORY_DESCRIPTIONS['cell_line_properties'],
+                                    path="/api")
 analysis_ns = Namespace("Analysis Queries", CATEGORY_DESCRIPTIONS['analysis'], path="/api")
 subnetwork_ns = Namespace("Subnetwork Queries", CATEGORY_DESCRIPTIONS['subnetwork'], path="/api")
 
@@ -320,7 +327,6 @@ def get_example_data():
     }
 }
 
-
 examples_dict = {
     "tissue": fields.List(fields.String, example=["UBERON", "UBERON:0001162"]),
     "gene": {
@@ -402,6 +408,9 @@ def get_example_data():
     "positive_genes": fields.List(fields.String, example=EXAMPLE_POSITIVE_HGNC_IDS),
     "negative_genes": fields.List(fields.String, example=EXAMPLE_NEGATIVE_HGNC_IDS),
     "gene_names": fields.List(fields.String, example=continuous_analysis_example_names),
+    "target_id": fields.String(example="hgnc:646"),
+    "is_downstream": fields.Boolean(example=False),
+    "minimum_evidence": fields.Float(example=2),
     "log_fold_change": fields.List(fields.Float, example=continuous_analysis_example_data),
     "species": fields.String(example="human"),
     "permutations": fields.Integer(example=100),

src/indra_cogex/apps/search/search.py

@@ -1,4 +1,5 @@
 import json
+from typing import List, Optional, Mapping, Tuple
 
 from flask import Blueprint, render_template, request, jsonify, redirect, url_for
 from flask_jwt_extended import jwt_required
@@ -8,12 +9,20 @@
 from wtforms.fields.simple import BooleanField
 from wtforms.validators import DataRequired
 
-from indra_cogex.apps.utils import render_statements, format_stmts
+from indra_cogex.apps.utils import render_statements
+from indra_cogex.client import Neo4jClient, autoclient
 from indra_cogex.client.queries import *
+from indra_cogex.representation import norm_id
+
 __all__ = ["search_blueprint"]
 
+from indra_cogex.client.queries import enrich_statements
+
+from indra_cogex.representation import indra_stmts_from_relations
+
 search_blueprint = Blueprint("search", __name__, url_prefix="/search")
 
+
 class SearchForm(FlaskForm):
     agent_name = StringField("Agent Name", validators=[DataRequired()])
     agent_role = StringField("Agent Role")
@@ -29,7 +38,7 @@ class SearchForm(FlaskForm):
     submit = SubmitField("Search")
 
 
-@search_blueprint.route("/", methods=['GET','POST'])
+@search_blueprint.route("/", methods=['GET', 'POST'])
 @jwt_required(optional=True)
 def search():
     stmt_types = {c.__name__ for c in get_all_descendants(Statement)}
@@ -105,9 +114,11 @@ def search():
         stmt_types_json=stmt_types_json,
     )
 
+
 from flask import current_app
 
-@search_blueprint.route("/gilda_ground", methods=["GET","POST"])
+
+@search_blueprint.route("/gilda_ground", methods=["GET", "POST"])
 @jwt_required(optional=True)
 def gilda_ground_endpoint():
     data = request.get_json()
@@ -122,6 +133,7 @@ def gilda_ground_endpoint():
     except Exception as e:
         return {"error": str(e)}, 500
 
+
 def gilda_ground(agent_text):
     try:
         from gilda.api import ground
@@ -131,4 +143,149 @@ def gilda_ground(agent_text):
         res = requests.post('http://grounding.indra.bio/ground', json={'text': agent_text})
         return res.json()
     except Exception as e:
-        return {"error": f"Grounding failed: {str(e)}"}
+        return {"error": f"Grounding failed: {str(e)}"}
+
+
+@autoclient()
+def get_ora_statements(
+    target_id: str,
+    genes: List[str],
+    minimum_belief: float = 0.0,
+    minimum_evidence: Optional[int] = None,
+    is_downstream: bool = False,
+    *,
+    client: Neo4jClient,
+) -> Tuple[List[Statement], Mapping[int, int]]:
+    """Get statements connecting input genes to target entity for ORA analysis.
+
+    Parameters
+    ----------
+    target_id : str
+        The ID of the target entity (e.g., 'GO:0006955', 'MESH:D007239')
+    genes : List[str]
+        List of gene IDs (e.g., ['HGNC:6019', 'HGNC:11876'])
+    minimum_belief : float
+        Minimum belief score for relationships
+    minimum_evidence : Optional[int]
+        Minimum number of evidences required for a statement to be included
+    is_downstream : bool
+        Whether this is a downstream analysis
+    client : Neo4jClient
+        The Neo4j client to use for querying
+
+    Returns
+    -------
+    :
+        A tuple containing:
+        - List of INDRA statements representing the relationships
+        - Dictionary mapping statement hashes to their evidence counts
+    """
+    # Normalize gene IDs
+    normalized_genes = [norm_id('HGNC', gene.split(':')[1]) for gene in genes]
+    print(f"DEBUG: Normalized genes: {normalized_genes[:5]}...")
+
+    # Handle different entity types and their relationships
+    namespace = target_id.split(':')[0].lower()
+    id_part = target_id.split(':')[1]
+
+    if namespace == 'mesh':
+        normalized_target = f"mesh:{id_part}"
+        rel_types = ["indra_rel", "has_indication"]
+    elif namespace == 'fplx':
+        normalized_target = f"fplx:{id_part}"
+        rel_types = ["indra_rel", "isa"]
+    else:
+        normalized_target = target_id.lower()
+        rel_types = ["indra_rel"]
+
+    # Main query for getting statements
+    query = """
+    MATCH p = (d:BioEntity {id: $target_id})-[r]->(u:BioEntity)
+    WHERE type(r) IN $rel_types
+    AND u.id STARTS WITH "hgnc"
+    AND NOT u.obsolete
+    AND u.id IN $genes
+    AND (type(r) <> 'indra_rel' OR r.belief > $minimum_belief)
+    WITH distinct r.stmt_hash AS hash, collect(p) as pp
+    RETURN pp
+    """
+
+    if is_downstream:
+        query = """
+        MATCH p = (u:BioEntity)-[r]->(d:BioEntity {id: $target_id})
+        WHERE type(r) IN $rel_types
+        AND u.id STARTS WITH "hgnc"
+        AND NOT u.obsolete
+        AND u.id IN $genes
+        AND (type(r) <> 'indra_rel' OR r.belief > $minimum_belief)
+        WITH distinct r.stmt_hash AS hash, collect(p) as pp
+        RETURN pp
+        """
+
+    params = {
+        "target_id": normalized_target,
+        "genes": normalized_genes,
+        "rel_types": rel_types,
+        "minimum_belief": minimum_belief
+    }
+    results = client.query_tx(query, **params)
+    flattened_rels = [client.neo4j_to_relation(i[0]) for rel in results for i in rel]
+
+    # Filter relations based on minimum_evidence
+    if minimum_evidence:
+        flattened_rels = [
+            rel for rel in flattened_rels
+            if rel.data.get("evidence_count", 0) >= minimum_evidence
+        ]
+
+    stmts = indra_stmts_from_relations(flattened_rels, deduplicate=True)
+
+    # Enrich statements with complete evidence (no limit)
+    stmts = enrich_statements(
+            stmts,
+            client=client
+    )
+
+    # Create evidence count mapping
+    evidence_counts = {
+            stmt.get_hash(): rel.data.get("evidence_count", 0)
+            for rel, stmt in zip(flattened_rels, stmts)
+    }
+
+    return stmts, evidence_counts
+
+
+@search_blueprint.route("/ora_statements/", methods=['GET'])
+@jwt_required(optional=True)
+def search_ora_statements():
+    """Endpoint to get INDRA statements connecting input genes to a target entity."""
+    target_id = request.args.get("target_id")
+    genes = request.args.getlist("genes")
+    is_downstream = request.args.get("is_downstream", "").lower() == "true"
+
+    try:
+        minimum_evidence = int(request.args.get("minimum_evidence") or 2)
+        minimum_belief = float(request.args.get("minimum_belief") or 0.0)
+    except (ValueError, TypeError):
+        return jsonify({"error": "Invalid parameter values"}), 400
+
+    if not target_id or not genes:
+        return jsonify({"error": "target_id and genes are required"}), 400
+
+    try:
+        statements, evidence_counts = get_ora_statements(
+            target_id=target_id,
+            genes=genes,
+            minimum_belief=minimum_belief,
+            minimum_evidence=minimum_evidence,
+            is_downstream=is_downstream
+        )
+
+        return render_statements(
+            stmts=statements,
+            evidence_count=evidence_counts
+        )
+
+    except Exception as e:
+        print(f"Error in get_ora_statements: {str(e)}")
+        return jsonify({"error": str(e)}), 500

src/indra_cogex/apps/templates/gene_analysis/discrete_results.html

@@ -52,8 +52,8 @@
             $("#table-wikipathways").DataTable(datatablesConf);
             $("#table-hpo").DataTable(datatablesConf);
             {% if indra_downstream_results is not none %}
-            $("#table-downstream").DataTable(datatablesConf);
-            $("#table-upstream").DataTable(datatablesConf);
+            $("#indra-table-downstream").DataTable(datatablesConf);
+            $("#indra-table-upstream").DataTable(datatablesConf);
             {% endif %}
         });
     </script>
@@ -82,6 +82,45 @@
     </table>
 {% endmacro %}
 
+{% macro render_indra_table(df, table_id) -%}
+   {% set is_downstream = "downstream" in table_id %}
+   <table class="table table-hover table-striped table-ora" id="indra-{{ table_id }}" style="width: 100%;">
+       <thead>
+       <tr>
+           <th scope="col">CURIE</th>
+           <th scope="col">Name</th>
+           <th scope="col"><i>p</i>-value</th>
+           <th scope="col"><i>q</i>-value</th>
+           <th scope="col">Statements</th>
+       </tr>
+       </thead>
+       <tbody>
+       {% for curie, name, p, mlp, q, mlq in df.values %}
+           {% set gene_list = [] %}
+           {% for gene_id in genes.keys() %}
+               {% do gene_list.append("HGNC:" + gene_id) %}
+           {% endfor %}
+           <tr>
+               <td><a href="https://bioregistry.io/{{ curie }}" target="_blank">{{ curie }}</a></td>
+               <td>{{ name }}</td>
+               <td>{{ "{:.2e}".format(p) }}</td>
+               <td>{{ "{:.2e}".format(q) }}</td>
+               <td>
+                   <a href="{{ url_for('search.search_ora_statements',
+                       target_id=curie,
+                       genes=gene_list,
+                       minimum_evidence=minimum_evidence,
+                       minimum_belief=minimum_belief,
+                       is_downstream=is_downstream) }}"
+                       target="_blank"
+                       class="btn btn-primary btn-sm">View Statements</a>
+               </td>
+           </tr>
+       {% endfor %}
+       </tbody>
+   </table>
+{% endmacro %}
+
 {% block container %}
     <div class="card card-body bg-light">
         <h1 class="display-3">Gene List Analysis</h1>
@@ -195,7 +234,7 @@ <h3>Over-Representation Analysis</h3>
                             </p>
                             <p>
                             </p>
-                            {{ render_table(indra_downstream_results, "table-downstream") }}
+                            {{ render_indra_table(indra_downstream_results, "table-downstream") }}
                         </div>
                         <div class="tab-pane" id="upstream" role="tabpanel" aria-labelledby="upstream-tab">
                             <p>
@@ -206,7 +245,7 @@ <h3>Over-Representation Analysis</h3>
                                 and correcting using {{ method }} and α={{ alpha }} on genes causally upstream in
                                 one step from all entities in the INDRA Database.
                             </p>
-                            {{ render_table(indra_upstream_results, "table-upstream") }}
+                            {{ render_indra_table(indra_upstream_results, "table-upstream") }}
                         </div>
                     {% endif %}
                 </div>