PlantGraph Knowledge Graph & API

Exceeding returns 429.

API Endpoints

Graph

Search

Enrichment

AI

Paper Evidence

Auth & Status

ID Format by Species

Critical: Arabidopsis uses l.name for lookups. All other species use l.id.

ID Conversion from Phytozome to PlantGraph

B. sylvaticum Strategy

Not in PlantGraph. Use orthologue chain:

1. Map Bsyl gene -> Bdis orthologue (Phytozome BioMart)
2. Convert Bdis ID to PlantGraph format (BRADI_{chr}g{num}v3)
3. Query PlantGraph for Bdis locus or follow ORTHOLOGOUS_TO to Arabidopsis

Key Node Labels

Gene-level: Locus, Transcript, Protein, Gene Annotation: GO, PO, ProteinDomain, InterProDomain, MapManBin, Pathway Taxonomy: Species, GeneFamily, PhytozomeGeneFamily, ProteinFamily Regulation: TranscriptionFactorFamily, Promoter, BindingSite, BindingMotif, TFMotif, RegulatoryElement Expression: Tissue, CellType, ExpressionProfile, DiurnalProfile, DevelopmentalStage Literature: Publication, Author, Journal, MeSH Variants: SNP, INDEL, GWASStudy, Trait, PlantTrait Other: UniProt, Phenotype, Hormone, Pathogen, Complex, Metabolite

Key Relationship Types (149 total)

Core biology: INTERACTS_WITH, COEXPRESSES_WITH, COFUNCTIONAL_WITH (AraNet), ORTHOLOGOUS_TO, PARALOGOUS_TO Annotation: ANNOTATED_WITH (GO), ANNOTATED_WITH_PO, HAS_DOMAIN, HAS_INTERPRO_DOMAIN, HAS_MAPMAN_BIN Regulation (outgoing): ACTIVATES, REPRESSES, REGULATES, BINDS_TO, BINDS_PROMOTER, TARGETS Regulation (incoming): REGULATES, DAPSEQ_BINDS_PROMOTER (gold-standard DAP-seq TF binding), TARGETS, ACTIVATES, REPRESSES Membership: MEMBER_OF, MEMBER_OF_TF_FAMILY, MEMBER_OF_GENE_FAMILY, MEMBER_OF_COMPLEX, PART_OF_COMPLEX, PARTICIPATES_IN Expression: EXPRESSED_IN, EXPRESSED_IN_TISSUE (target is DevelopmentalStage, not Tissue), EXPRESSED_IN_CELL_TYPE, EXPRESSED_UNDER_STRESS Responses: RESPONDS_TO_HORMONE, RESPONDS_TO_NUTRIENT (N, P, K, Fe, Ca, S, Zn), RESPONDS_TO_PATHOGEN Life history: INVOLVED_IN_FLOWERING (1,479 genes), CITED_IN_FLOWERING, INVOLVED_IN_ADAPTATION (vernalisation, photoperiod, thermosensing, drought, heat) GWAS/traits: ASSOCIATED_WITH_TRAIT (flowering_time, biomass, root_length, seed_yield, stress_response) Epigenetics: HAS_HISTONE_MARK (H3K27me3, H3K4me3, H4K16ac, etc.), HAS_METHYLATION PTMs: HAS_PHOSPHORYLATION_SITE, HAS_UBIQUITINATION_SITE, HAS_ACETYLATION_SITE, HAS_SUMOYLATION_SITE, HAS_SNITROSYLATION_SITE, HAS_SULFENYLATION_SITE Literature: PUBLISHED_IN, MENTIONED_IN, CITED_IN_FLOWERING Pathway/enzyme: BELONGS_TO_PATHWAY, CATALYZES, CATALYZES_REACTION, ENCODES_ENZYME

Lookup Methods

1. By ID (primary)

-- Arabidopsis (has both l.name and l.id set)
MATCH (l:Locus {name: "AT2G40150"}) RETURN l.name, l.symbol, l.description

-- Non-Arabidopsis (l.name is null, use l.id)
MATCH (l:Locus {id: "BRADI_1g16610v3"}) RETURN l.id, l.species, properties(l)

2. By gene symbol

MATCH (l:Locus {symbol: "TBL28"}) RETURN l.id, l.species
-- Pattern match:
MATCH (l:Locus) WHERE l.symbol CONTAINS "TBL" RETURN l.id, l.symbol LIMIT 20

Only Ath and some rice loci have symbols.

3. By fulltext search (cross-species)

CALL db.index.fulltext.queryNodes("gene_fulltext", "trichome birefringence")
YIELD node, score
RETURN node.id, node.symbol, node.species, score LIMIT 10

Indexes: gene_fulltext, go_fulltext, publication_fulltext, pathway_fulltext, compound_fulltext, enzyme_fulltext, reaction_fulltext.

4. By UniProt accession (reverse lookup)

MATCH (l:Locus)-[:MAPS_TO]->(u:UniProt {id: "Q94K00"}) RETURN l.id, l.name, l.species

5. Via /search/terms (Arabidopsis only)

GET /api/v1/search/terms?q=TBL28      -> {"term":"TBL28","id":"AT2G40150","type":"gene"}
GET /api/v1/search/terms?q=AT2G40150   -> same

Resolves Ath symbols and TAIR IDs only.

Top 10 Common Query Patterns

1. Get GO annotations

MATCH (l:Locus {name: "AT2G40150"})-[:ANNOTATED_WITH]->(go:GO)
RETURN go.id, go.name

2. Get protein interactions

MATCH (l:Locus {name: "AT2G40150"})-[:INTERACTS_WITH]->(p:Locus)
RETURN p.name, p.symbol, p.description

3. Get co-expression partners

MATCH (l:Locus {name: "AT2G40150"})-[:COEXPRESSES_WITH]->(c:Locus)
RETURN c.name, c.symbol, c.description

4. Get regulatory network (incoming edges)

MATCH (tf:Locus)-[r:REGULATES|ACTIVATES|REPRESSES]->(l:Locus {name: "AT2G40150"})
RETURN tf.name, tf.symbol, type(r) AS regulation_type

5. Get orthologues (specific species)

Note: ORTHOLOGOUS_TO direction is arbitrary. Always use undirected match for cross-species lookups.

MATCH (a:Locus {name: "AT2G40150"})-[:ORTHOLOGOUS_TO]->(o:Locus)
WHERE o.species = "Brachypodium distachyon"
RETURN o.id, o.species

6. Get orthologues (all species with counts)

MATCH (a:Locus {name: "AT2G40150"})-[:ORTHOLOGOUS_TO]->(o:Locus)
WHERE o.species IS NOT NULL
RETURN o.species, count(*) AS n ORDER BY n DESC

7. Find Ath orthologue from non-Ath gene

Critical: Use undirected match (no arrow):

MATCH (b:Locus {id: "BRADI_1g16610v3"})-[:ORTHOLOGOUS_TO]-(a:Locus)
WHERE a.species = "Arabidopsis thaliana"
RETURN a.name, a.symbol, a.description

8. Multi-hop: Brachypodium -> Arabidopsis -> interactions

MATCH (b:Locus {id: "BRADI_1g16610v3"})-[:ORTHOLOGOUS_TO]-(a:Locus)
WHERE a.species = "Arabidopsis thaliana"
WITH a
MATCH (a)-[:INTERACTS_WITH]->(p:Locus)
RETURN a.name AS ath_gene, p.name AS interactor, p.description

9. Bulk query: multiple genes

MATCH (l:Locus)
WHERE l.name IN ["AT1G01430","AT2G40150","AT5G06700"]
OPTIONAL MATCH (l)-[:ANNOTATED_WITH]->(go:GO)
RETURN l.name, l.symbol, collect(DISTINCT go.name) AS go_terms

10. Schema introspection

CALL db.labels() YIELD label RETURN label
CALL db.relationshipTypes() YIELD relationshipType RETURN relationshipType
MATCH (l:Locus) RETURN DISTINCT l.species, count(*) AS n ORDER BY n DESC

Graph API (REST)

All /api/graph/ require JWT (-H "Authorization: Bearer $TOKEN").

GET examples

curl -s https://plantgraph.se/api/graph/gene/AT2G40150 -H "Authorization: Bearer $TOKEN"
curl -s "https://plantgraph.se/api/graph/search?q=trichome&limit=10" -H "Authorization: Bearer $TOKEN"
curl -s "https://plantgraph.se/api/graph/gene/AT2G40150/neighbors?rel_type=INTERACTS_WITH&direction=both" -H "Authorization: Bearer $TOKEN"
curl -s https://plantgraph.se/api/graph/gene/AT2G40150/pathways -H "Authorization: Bearer $TOKEN"
curl -s https://plantgraph.se/api/graph/gene/AT2G40150/stats -H "Authorization: Bearer $TOKEN"
curl -s https://plantgraph.se/api/graph/pathway/PWY-5136 -H "Authorization: Bearer $TOKEN"
curl -s "https://plantgraph.se/api/graph/pathways?source=KEGG&q=photosynthesis" -H "Authorization: Bearer $TOKEN"
curl -s https://plantgraph.se/api/graph/stats -H "Authorization: Bearer $TOKEN"

POST examples

# Interaction network
curl -s -X POST https://plantgraph.se/api/graph/network/interactions \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"gene_ids": ["AT2G40150", "AT1G01430", "AT5G06700"]}'

# Co-expression network
curl -s -X POST https://plantgraph.se/api/graph/network/coexpression \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"gene_ids": ["AT2G40150", "AT1G01430"], "min_score": 0.7}'

# Regulatory network (TF -> target)
curl -s -X POST https://plantgraph.se/api/graph/network/regulatory \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"gene_ids": ["AT2G40150"], "include_dapseq": true}'

# NL query -> Cypher + results + confidence
curl -s -X POST https://plantgraph.se/api/graph/query/natural \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"question": "What genes interact with FLC in Arabidopsis?"}'

# Raw Cypher (authenticated)
curl -s -X POST https://plantgraph.se/api/graph/query \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"query": "MATCH (l:Locus {name: \"AT2G40150\"}) RETURN l LIMIT 1"}'

No-auth endpoints

# Raw Cypher
curl -s -X POST https://plantgraph.se/api/v1/graph/query \
  -H "Content-Type: application/json" \
  -d '{"query":"MATCH (l:Locus {name: \"AT2G40150\"}) RETURN l.name, l.description LIMIT 1"}'

# Multi-species gene search
curl -s "https://plantgraph.se/api/v1/search/genes?q=flowering&limit=20&offset=0"

Enrichment API

/api/enrichment/ requires JWT unless noted.

Query patterns

# All 42 sources (~30s)
curl -s https://plantgraph.se/api/enrichment/gene/AT2G40150 -H "Authorization: Bearer $TOKEN"
# Selective sources
curl -s "https://plantgraph.se/api/enrichment/gene/AT2G40150?sources=uniprot,string,alphafold" -H "Authorization: Bearer $TOKEN"
# Single source
curl -s https://plantgraph.se/api/enrichment/gene/AT2G40150/uniprot -H "Authorization: Bearer $TOKEN"

42 Enrichment Sources

Protein & Structure (7): uniprot (function, PTMs, sequence), alphafold (3D structure, pLDDT), interpro (domains, sites), panther (families, orthologs), pdbe (experimental structures), rcsb_pdb (structures + quality), ebi_proteins (features, variants)

Interactions (3): string (PPI scores, score_threshold param, default 0.4), intact (experimental interactions), biogrid (physical + genetic)

Pathways (3): kegg (500+ Ath pathways), reactome (curated reactions), wikipathways (community, 200+ species)

Expression (3): expression_atlas (3000+ experiments, experiment_type param), bar_efp (tissue expression), atted2 (co-expression scores)

Ontology (3): quickgo (GO + evidence codes), chebi (chemicals, SMILES/InChI), ols (200+ EBI ontologies)

Literature (3): europepmc (40M+ articles, sort param), semantic_scholar (200M+ papers), crossref (130M+ DOIs)

Comparative (4): ensembl (orthologs, gene trees, 7 plant spp), oma (2100+ genomes), gramene (66 plant spp), plaza (families, synteny, 50+ genomes)

Plant-specific (6): planttfdb (TF classification, 2000+ Ath TFs), suba (subcellular localization), aranet (co-functional, 22K+ genes), arapheno (QTL, 1135 accessions), thalemine (integrated info), 1001genomes (SNPs/indels, 1135 accessions)

Other (4): rnacentral (ncRNA, 20M+ seq), pubchem (100M+ compounds), wikidata (linked data), mygene (annotation aggregation)

Batch, GO, Pathway enrichment

# Batch (multiple genes + sources)
curl -s -X POST https://plantgraph.se/api/enrichment/batch \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"genes":["AT2G40150","AT1G01430","AT5G06700"],"sources":["uniprot","string"],"parallel":true}'

# GO enrichment (returns enriched terms with p-values, fold enrichment)
curl -s -X POST https://plantgraph.se/api/enrichment/analysis/go \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"genes":["AT2G40150","AT1G01430","AT5G06700","AT3G20810"],"background":"genome","ontology":"BP","correction":"fdr"}'

# Pathway enrichment
curl -s -X POST https://plantgraph.se/api/enrichment/analysis/pathway \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"genes":["AT2G40150","AT1G01430","AT5G06700"],"databases":["kegg","reactome"],"p_threshold":0.05}'

# Legacy (Ath only, no auth, min 2 genes, p_threshold max 0.5)
curl -s -X POST https://plantgraph.se/api/v1/gene-enrichment/analyze \
  -H "Content-Type: application/json" \
  -d '{"gene_ids":["AT1G01010","AT1G01020"],"categories":["GO"],"p_threshold":0.05}'

# Cache status
curl -s https://plantgraph.se/api/enrichment/cache/status/AT2G40150 -H "Authorization: Bearer $TOKEN"

AI API

/api/ AI endpoints require JWT.

Questions & Chat

# Generate questions from abstract
curl -s -X POST https://plantgraph.se/api/questions/generate-from-abstract \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"abstract": "We investigated the role of FLC in vernalization-mediated flowering...", "num_questions": 5}'

# From PMID
curl -s -X POST https://plantgraph.se/api/questions/generate-from-pmid \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"pmid": "34567890"}'

# Random PMID from 290K publications
curl -s https://plantgraph.se/api/questions/random-pmid -H "Authorization: Bearer $TOKEN"

# Chat (with sources)
curl -s -X POST https://plantgraph.se/api/chat/message \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"session_id": "sess_abc123", "message": "What are the main regulators of flowering time in Arabidopsis?"}'

# Session: create / list / history
curl -s -X POST https://plantgraph.se/api/chat/sessions \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"name": "Flowering research"}'
curl -s https://plantgraph.se/api/chat/sessions -H "Authorization: Bearer $TOKEN"
curl -s https://plantgraph.se/api/chat/sessions/sess_abc123/history -H "Authorization: Bearer $TOKEN"

AI Query, Hypothesis, GNN

# NL -> Cypher + results + confidence
curl -s -X POST https://plantgraph.se/api/ai/query \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"question": "Which transcription factors regulate FLC expression?"}'

# Hypothesis generation
curl -s -X POST https://plantgraph.se/api/ai/hypothesis \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"gene_ids": ["AT5G10140", "AT3G18990"], "context": "cold stress response"}'

# GNN function predictions
curl -s https://plantgraph.se/api/gnn/predict/AT2G40150 -H "Authorization: Bearer $TOKEN"

# GNN link prediction (novel relationships with probability)
curl -s -X POST https://plantgraph.se/api/gnn/predict/links \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"gene_id": "AT2G40150", "rel_types": ["INTERACTS_WITH", "COEXPRESSES_WITH"], "top_k": 20}'

Q&A Discovery (autonomous loop)

# Start
curl -s -X POST https://plantgraph.se/api/qa-discovery/start \
  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
  -d '{"topic": "cell wall biosynthesis in grasses", "max_iterations": 10, "depth": "deep"}'

# Check progress / Get results
curl -s https://plantgraph.se/api/qa-discovery/disc_xyz789/status -H "Authorization: Bearer $TOKEN"
curl -s https://plantgraph.se/api/qa-discovery/disc_xyz789/results -H "Authorization: Bearer $TOKEN"

Paper Evidence API

/api/v1/ prefix, no auth. Claim verification takes 30-70s.

Check abstract (score 0-100)

curl -s -X POST https://plantgraph.se/api/v1/paper-evidence/check-abstract \
  -H "Content-Type: application/json" \
  -d '{"abstract_text": "FLC is a key repressor of flowering in Arabidopsis, regulated by vernalization through epigenetic silencing via Polycomb..."}'
# Returns: {overall_score, score_tier, entities[], evidence{graph_support, literature_support, interaction_support}}

Verify claims (SSE stream)

curl -s -X POST https://plantgraph.se/api/v1/paper-evidence/verify-claims \
  -H "Content-Type: application/json" \
  -d '{"abstract_text": "FLC is repressed by VIN3 during vernalization...", "claims": ["FLC interacts with SVP to repress flowering", "VIN3 is induced by cold exposure"]}'
# SSE events: claim_extracted -> entity_report -> claim_verified (verdict + confidence) -> complete

If claims omitted, auto-extracted from abstract. Verdicts: supported, partially_supported, not_found.

PDF endpoints

curl -s -X POST https://plantgraph.se/api/v1/paper-evidence/check -F "[email protected]"           # Score
curl -s -X POST https://plantgraph.se/api/v1/paper-evidence/verify-claims-pdf -F "[email protected]" # SSE stream

Cypher Deep Reference

Node Labels (132)

Gene-level: Locus, Transcript, Protein, Gene, SmallORF/sORF, CircRNA/CircularRNA, LncRNA/lncRNA, SmallRNA

Annotation: GO, PO, TO, BTO, PECO, ENVO, ProteinDomain, InterProDomain, MapManBin, Pathway, Enzyme, Reaction, MetabolicCompound

Taxonomy: Species, NCBITaxon, GeneFamily, PhytozomeGeneFamily, ProteinFamily

Regulation: TranscriptionFactorFamily, Promoter, BindingSite, BindingMotif, TFMotif, Motif, RegulatoryElement, ReMapPeak, PeakRegion, DHSRegion, ChromatinAccessibilityPeak, SignalTrack

Expression: Tissue, CellType, ExpressionProfile (7.3M), DiurnalProfile, DevelopmentalStage, SpatialRegion, SpatialZone, DiurnalCondition

Experiment: Experiment, Study, Sample, GrowthCondition, Condition, StressCondition, StressExperiment, StressTimePoint, TimePoint, RnaSeqLibrary, MicroarrayPlatform, EBIAtlasExperiment, BARExperiment, GSE, GSM, MetaboLightsStudy

Literature: Publication (290K), Preprint, Author (105K), Journal, MeSH

Variants/GWAS: SNP (1.3M), INDEL, GWASStudy, Trait, PlantTrait, AraphenoStudy, Germplasm

Protein features: UniProt, ProteinComplex, Complex, ProteinStructure, ProteinFeature, ProteinLigand, Ligand, LigandBinding, LigandBindingSite, ProteomicsEvidence

PTM sites: PhosphorylationSite (190K), UbiquitinationSite, AcetylationSite, SUMOylationSite, SnitrosylationSite, SulfenylationSite, GlcNAcylationSite, MethylationSite (2.9M), PTMSite

Chemistry: ChEBICompound, KnapsackCompound, Metabolite, NaturalProduct, Phytochemical, PhytochemicalCompound, PhytohubCompound

Epigenetics: MethylationContext, HistoneModification (154K)

Pathogen: Pathogen, PathogenEffector, PathogenGene, PHIGene, PHIInteraction, PHIOrganism, PHIPOTerm, MicrobeInteraction

Biosynthesis: BGC_Cluster, BGCluster, BiosynthesisCluster, BiosyntheticCluster (legacy duplicates)

Other: SubcellularCompartment, CellularCompartment, GeneXref, OntologyAnchor, Hormone, Nutrient, Phenotype, ClimateAdaptation, Institution, PMNDatabase, TAIRLocus, TAIRObject

Relationship Types (149)

Top 20 by count:

LIMIT aggressively on high-count rels. INTERACTS_WITH/COEXPRESSES_WITH can return 100K+ rows.

Full list by category:

Core biology: INTERACTS_WITH, COEXPRESSES_WITH, COFUNCTIONAL_WITH, ORTHOLOGOUS_TO, PARALOGOUS_TO

Annotation: ANNOTATED_WITH, ANNOTATED_WITH_PO, NCBI_ANNOTATED_WITH, HAS_DOMAIN, HAS_INTERPRO_DOMAIN, HAS_PROTEIN_DOMAIN, HAS_MAPMAN_BIN, DOMAIN_GO_ANNOTATION

Regulation: ACTIVATES, REPRESSES, REGULATES, BINDS_TO, BINDS_PROMOTER, DAPSEQ_BINDS_PROMOTER, TARGETS, HAS_BINDING_MOTIF, HAS_BINDING_SITE, HAS_MOTIF, HAS_JASPAR_MOTIF, HAS_REPRESENTATIVE_MOTIF, BINDS_AT

Membership: MEMBER_OF, MEMBER_OF_TF_FAMILY, MEMBER_OF_GENE_FAMILY, MEMBER_OF_COMPLEX, MEMBER_OF_CLUSTER, PART_OF_COMPLEX, PART_OF_CLUSTER, PARTICIPATES_IN, BELONGS_TO_PATHWAY, HAS_SUBPATHWAY, IN_CLUSTER, PARENT_BIN

Expression: EXPRESSED_IN, EXPRESSED_IN_TISSUE, EXPRESSED_IN_CELL_TYPE, EXPRESSED_IN_REGION, EXPRESSED_IN_ZONE, EXPRESSED_UNDER_STRESS, EXPRESSED_UNDER_CONDITION, EXPRESSED_AT_TIMEPOINT, HAS_PROFILE, HAS_DIURNAL_PROFILE, HAS_RHYTHMIC_MEMBERS, REPRESENTS_STAGE

Spatial: LOCATED_IN_SPATIAL_REGION, HAS_SPATIAL_LOCATION, LOCATED_IN, LOCALIZED_TO, LOCALIZES_TO

Responses: RESPONDS_TO_HORMONE, RESPONDS_TO_NUTRIENT, RESPONDS_TO_PATHOGEN, AFFECTED_BY_CONDITION, UNDER_CONDITION, HAS_CONDITION

Life history: INVOLVED_IN_FLOWERING, CITED_IN_FLOWERING, INVOLVED_IN_ADAPTATION, INVOLVED_IN_SYMBIOSIS

GWAS/traits: ASSOCIATED_WITH_TRAIT, ASSOCIATED_WITH, IS_A_TRAIT, MEASURES_TRAIT, GWAS_OF, HAS_PHENOTYPE

Epigenetics: HAS_HISTONE_MARK, HAS_METHYLATION, METHYLATED_IN_CONTEXT, HAS_METHYLATION_SITE, HAS_DHS_REGION, IN_ACCESSIBLE_REGION

PTMs: HAS_PHOSPHORYLATION_SITE, HAS_PHOSPHOSITE, HAS_UBIQUITINATION_SITE, HAS_ACETYLATION_SITE, HAS_SUMOYLATION_SITE, HAS_SNITROSYLATION_SITE, HAS_SULFENYLATION_SITE, HAS_GLCNACYLATION_SITE, HAS_MYRISTOYLATION, HAS_PTM_SITE

Protein: MAPS_TO, ENCODES, HAS_FEATURE, HAS_STRUCTURE, BINDS_LIGAND, HAS_LIGAND_BINDING, HAS_MS_EVIDENCE, HAS_PROTEOMICS_EVIDENCE, TRANSLATED_FROM, EXPRESSED_AS, HAS_TRANSCRIPT, ENCODES_ENZYME, ENCODES_LNCRNA, ENCODES_SORF, HAS_SMALL_ORF

Pathogen: HAS_PATHOGEN, TARGETS_HOST_PROTEIN, VIRULENCE_GENE_OF, HAS_HOST, FOUND_IN_ORGANISM, INVOLVES_GENE

Metabolism: CATALYZES, CATALYZES_REACTION, HAS_REACTION, CONSUMES, PRODUCES, METABOLITE_IN

Literature: PUBLISHED_IN, MENTIONED_IN, CITED_BY, CITES, CITED_IN_FLOWERING, HAS_AUTHOR, COAUTHORED_WITH, HAS_MESH_TERM, LATER_PUBLISHED_AS, WORKS_ON_GENE

Variant: HAS_VARIANT, NEAR_VARIANT, OVERLAPS_WITH

Cross-reference: SAME_AS, DERIVED_FROM, DERIVES_FROM, CONNECTED_TO_GRAPH, HAS_TAIR_LOCUS, HAS_TAIR_OBJECT, ONTOLOGY_ANCHOR, FROM_DATABASE, SUPPORTED_BY, DETECTED_BY, MARKER_OF, MEASURED, MEASURED_IN, MEASURED_UNDER, OBSERVED_IN, HAS_EXPERIMENT, PART_OF_EXPERIMENT, HAS_SAMPLE, HAS_PECO_TERM, HAS_TO_TERM, IN_TAXON, PARENT_OF, IS_A

Relationship Properties

Filter GO to experimental evidence:

MATCH (l:Locus {name: "AT2G40150"})-[r:ANNOTATED_WITH]->(go:GO)
WHERE r.evidence IN ["IDA", "IMP", "IGI", "IPI", "IEP"]
RETURN go.id, go.name, r.evidence

Filter co-expression by score:

MATCH (l:Locus {name: "AT2G40150"})-[r:COEXPRESSES_WITH]->(c:Locus)
WHERE r.correlation_score > 0.8
RETURN c.name, c.symbol, r.correlation_score
ORDER BY r.correlation_score DESC LIMIT 20

Filter stress expression by log2fc:

MATCH (l:Locus {name: "AT2G40150"})-[r:EXPRESSED_UNDER_STRESS]->(s)
WHERE r.log2fc > 1
RETURN s.name AS stress, r.log2fc
ORDER BY r.log2fc DESC

Node Properties

Locus (Arabidopsis): id, name (AT ID), symbol, species, taxon_id, chromosome, start, end, strand, description, gene_type, is_tf, tf_family, is_arabidopsis (true), ncbi_gene_id, entrez_id, uniprot_id, pfam_ids (array), subcellular_location, peptideatlas_detected, ensembl_homolog_count, aranet_enriched, remap_peak_count, remap_tf_count. Embedded expression: atgen_drought_*/atgen_salt_*/atgen_cold_* (ATGenExpress TPMs), circ_emexp1304_* (circadian), geod*_tpm/geod*_lfc, mtab*_tpm.

Locus (non-Arabidopsis): id, species, taxon_id, description, description_source, source, is_arabidopsis (false), ortholog_ref_id. Expression species-prefixed: bradi_mtab4401_*_tpm, bradi_mean_tpm, bradi_max_tpm. name is null -- use id.

GO: id, name, name_obo, namespace (molecular_function/biological_process/cellular_component), definition, synonyms (array), is_obsolete, enriched

Publication: pubmed_id, title, abstract, authors, year, journal, doi, pmc_id, citation_count, cited_by_count, is_open_access, full_text_mined, gene_count. Warning: research_concepts type inconsistent -- sometimes List[str], sometimes semicolon-delimited string.

Species: id (snake_case), name, common_name, taxon_id, source

Pathway: id, name, kegg_name, kegg_category, kegg_organism, kegg_gene_count, kegg_compound_count, url

Fulltext Indexes (7)

CALL db.index.fulltext.queryNodes("gene_fulltext", "cellulose synthase")
YIELD node, score
RETURN node.id, node.symbol, node.species, node.description, score
ORDER BY score DESC LIMIT 15

Key Indexes

Node: Locus.id (unique), Locus.name, Locus.species, Locus.symbol, Locus.is_tf, Locus.tf_family, Locus.taxon_id, Locus.entrez_id, Locus.ncbi_gene_id, Locus.chromosome+start+end (composite). Relationship: COEXPRESSES_WITH.correlation_score, EXPRESSED_UNDER_STRESS.log2fc, INTERACTS_WITH.source, DAPSEQ_BINDS_PROMOTER.source.

Additional Query Patterns

-- DAP-seq TF binding (gold standard)
MATCH (tf:Locus)-[:DAPSEQ_BINDS_PROMOTER]->(l:Locus {name: "AT2G40150"})
RETURN tf.name, tf.symbol

-- GWAS trait associations
MATCH (l:Locus {name: "AT2G40150"})-[:ASSOCIATED_WITH_TRAIT]->(t)
RETURN t.name AS trait

-- Protein complex membership
MATCH (l:Locus {name: "AT2G40150"})-[:MEMBER_OF_COMPLEX|PART_OF_COMPLEX]->(c)
RETURN c.name AS complex_name

-- Nutrient responses
MATCH (l:Locus {name: "AT1G02860"})-[:RESPONDS_TO_NUTRIENT]->(n)
RETURN n.name AS nutrient

-- Pathogen responses
MATCH (l:Locus {name: "AT1G01190"})-[:RESPONDS_TO_PATHOGEN]->(p)
RETURN p.name AS pathogen

-- Publications mentioning gene
MATCH (l:Locus {name: "AT2G40150"})-[:PUBLISHED_IN|MENTIONED_IN]->(pub:Publication)
RETURN pub.title, pub.pmid LIMIT 10

-- Histone marks (h.id contains mark type, e.g. "AraENC_H3K27me3_Chr3_...")
MATCH (l:Locus {name: "AT3G20810"})-[:HAS_HISTONE_MARK]->(h)
RETURN h.id LIMIT 10

-- Cell-type expression (single-cell)
MATCH (l:Locus {name: "AT2G40150"})-[:EXPRESSED_IN_CELL_TYPE]->(ct:CellType)
RETURN ct.name LIMIT 10

-- Spatial expression
MATCH (l:Locus {name: "AT2G40150"})-[:EXPRESSED_IN_REGION]->(sr:SpatialRegion)
RETURN sr.name LIMIT 10

-- Methylation + context
MATCH (l:Locus {name: "AT3G20810"})-[:HAS_METHYLATION]->(m)
RETURN m.id LIMIT 10
MATCH (l:Locus {name: "AT3G20810"})-[:METHYLATED_IN_CONTEXT]->(mc:MethylationContext)
RETURN mc.id LIMIT 10

-- All-annotations dump
MATCH (l:Locus {name: "AT2G40150"})-[r]->(t)
RETURN type(r) AS rel, labels(t)[0] AS target_type,
       coalesce(t.name, t.id) AS target, t.description
LIMIT 50

Species in Graph

127 species, 1.62M loci. Top 20 by locus count:

Also: 9 Oryza spp, conifers (Picea abies), mosses (Physcomitrium patens), ferns (Selaginella moellendorffii), algae (Galdieria sulphuraria, Chara braunii), liverworts (Marchantia polymorpha), hornworts (Anthoceros angustus).

Not in graph: Hordeum vulgare, B. sylvaticum. Use orthologue chains via Bdis or Ath.

Rice note: Appears as both "Oryza sativa" (13K) and "Oryza sativa subsp. japonica" (22K). Check both.

Cypher Best Practices

• NULL on sorts: WHERE prop IS NOT NULL or ORDER BY prop NULLS LAST
• count{} not size(): WHERE count{(l)-[:ANNOTATED_WITH]->()} > 3
• EXISTS{}: WHERE EXISTS {MATCH (l)-[:INTERACTS_WITH]->()} (not deprecated exists())
• Explicit grouping: WITH clauses before aggregations
• elementId() not id(): id() is deprecated
• coalesce(l.name, l.id): handles species where l.name is null
• Always LIMIT: 14.5M+ nodes; unbounded queries are slow
• Filter ANNOTATED_WITH: WHERE r.evidence IN ["IDA","IPI","IMP","IGI","IEP"] (exclude IEA floods)
• Filter COEXPRESSES_WITH: WHERE r.correlation_score > 0.7 (3.8M edges)

Query checklist

1. No deprecated functions (id(), size(), exists())
2. NULL filters on sorted properties
3. LIMIT on exploratory queries
4. Undirected match for reverse orthologue lookups
5. coalesce(t.name, t.id) for cross-species results

Common Mistakes

• l.name for non-Ath species -- use l.id
• Phytozome Bradi1g* format -- convert to BRADI_1g*v3
• MSU rice IDs LOC_Os* -- use RAP format Os*
• Expecting barley HORVU.* -- not in graph; use orthologue hop
• Directed -> for reverse orthologue lookups -- use undirected -[:ORTHOLOGOUS_TO]-
• Wrong regulatory direction -- edges point FROM regulator TO target: (tf)-[:REGULATES]->(l {name: "X"})
• EXPRESSED_IN_TISSUE with (t:Tissue) -- target is DevelopmentalStage, use (t) unlabeled
• Missing LIMIT -- INTERACTS_WITH/COEXPRESSES_WITH can return 100K+ rows
• Rice species string -- check both "Oryza sativa" (13K) and "Oryza sativa subsp. japonica" (22K)
• /api/ vs /api/v1/ confusion -- /api/ needs JWT; /api/v1/ mostly no auth
• Enrichment for non-Ath genes -- support varies by source
• Unfiltered ANNOTATED_WITH -- IEA floods results; filter by evidence code
• Unfiltered COEXPRESSES_WITH -- 3.8M edges; add correlation_score > 0.7
• Publication.research_concepts type varies -- sometimes List[str], sometimes semicolon-delimited string; handle both

R Integration

library(httr)
library(jsonlite)

plantgraph_login <- function(username, password) {
  resp <- POST(
    "https://plantgraph.se/api/auth/login",
    body = list(username = username, password = password),
    encode = "json",
    content_type_json()
  )
  content(resp, as = "parsed")$access_token
}

plantgraph_query <- function(cypher, token = NULL) {
  if (is.null(token)) {
    url <- "https://plantgraph.se/api/v1/graph/query"
    resp <- POST(url, body = list(query = cypher), encode = "json", content_type_json())
  } else {
    url <- "https://plantgraph.se/api/graph/query"
    resp <- POST(url, body = list(query = cypher), encode = "json",
                 content_type_json(), add_headers(Authorization = paste("Bearer", token)))
  }
  content(resp, as = "parsed", simplifyVector = TRUE)$results
}

plantgraph_gene <- function(gene_id, token) {
  resp <- GET(
    paste0("https://plantgraph.se/api/graph/gene/", gene_id),
    add_headers(Authorization = paste("Bearer", token))
  )
  content(resp, as = "parsed", simplifyVector = TRUE)
}

plantgraph_enrich <- function(gene_id, sources = NULL, token) {
  url <- paste0("https://plantgraph.se/api/enrichment/gene/", gene_id)
  if (!is.null(sources)) url <- paste0(url, "?sources=", paste(sources, collapse = ","))
  resp <- GET(url, add_headers(Authorization = paste("Bearer", token)))
  content(resp, as = "parsed", simplifyVector = TRUE)
}

plantgraph_go_enrichment <- function(gene_ids, p_threshold = 0.05, token) {
  resp <- POST(
    "https://plantgraph.se/api/enrichment/analysis/go",
    body = list(gene_ids = gene_ids, categories = list("biological_process"),
                p_threshold = p_threshold, correction = "benjamini_hochberg"),
    encode = "json",
    content_type_json(),
    add_headers(Authorization = paste("Bearer", token))
  )
  content(resp, as = "parsed", simplifyVector = TRUE)
}

# Example: Brachypodium GO terms (no auth)
plantgraph_query('
  MATCH (l:Locus {id: "BRADI_1g16610v3"})-[:ANNOTATED_WITH]->(go:GO)
  RETURN go.id, go.name
')

# token <- plantgraph_login("user", "pass")
# plantgraph_enrich("AT2G40150", sources = c("uniprot", "string"), token = token)

Python Integration

import requests

BASE_URL = "https://plantgraph.se"

def plantgraph_login(username: str, password: str) -> str:
    """Return JWT token."""
    resp = requests.post(
        f"{BASE_URL}/api/auth/login",
        json={"username": username, "password": password}
    )
    resp.raise_for_status()
    return resp.json()["access_token"]


def plantgraph_query(cypher: str, token: str | None = None) -> list:
    """Run Cypher. Uses /api/v1/ (no auth) if token is None."""
    if token is None:
        url = f"{BASE_URL}/api/v1/graph/query"
        headers = {"Content-Type": "application/json"}
    else:
        url = f"{BASE_URL}/api/graph/query"
        headers = {
            "Content-Type": "application/json",
            "Authorization": f"Bearer {token}",
        }
    resp = requests.post(url, json={"query": cypher}, headers=headers)
    resp.raise_for_status()
    return resp.json().get("results", [])


def plantgraph_gene(gene_id: str, token: str) -> dict:
    """Get gene details."""
    resp = requests.get(
        f"{BASE_URL}/api/graph/gene/{gene_id}",
        headers={"Authorization": f"Bearer {token}"}
    )
    resp.raise_for_status()
    return resp.json()


def plantgraph_enrich(gene_id: str, token: str, sources: list[str] | None = None) -> dict:
    """Enrich gene from external sources."""
    url = f"{BASE_URL}/api/enrichment/gene/{gene_id}"
    if sources:
        url += f"?sources={','.join(sources)}"
    resp = requests.get(url, headers={"Authorization": f"Bearer {token}"})
    resp.raise_for_status()
    return resp.json()


def plantgraph_go_enrichment(
    gene_ids: list[str], token: str, p_threshold: float = 0.05
) -> dict:
    """GO enrichment on gene list."""
    resp = requests.post(
        f"{BASE_URL}/api/enrichment/analysis/go",
        json={
            "gene_ids": gene_ids,
            "categories": ["biological_process"],
            "p_threshold": p_threshold,
            "correction": "benjamini_hochberg",
        },
        headers={
            "Content-Type": "application/json",
            "Authorization": f"Bearer {token}",
        },
    )
    resp.raise_for_status()
    return resp.json()


# No-auth example
results = plantgraph_query('''
    MATCH (l:Locus {name: "AT2G40150"})-[:ANNOTATED_WITH]->(go:GO)
    RETURN go.id, go.name
''')

# token = plantgraph_login("user", "pass")
# gene = plantgraph_gene("AT2G40150", token)
# enrichment = plantgraph_enrich("AT2G40150", token, sources=["uniprot", "string"])
# go_results = plantgraph_go_enrichment(["AT2G40150", "AT1G01430", "AT5G06700"], token)