EFNB1

Summary: Association with the signaling receptor ERBB2 is more informative than generic protein binding.

Reason: The human EFNB1–ERBB2 association is supported by endogenous co-immunoprecipitation in human epithelial and breast cancer cell lines and by tagged-construct experiments in PMID:22279592. The exact pair is independently represented in IntAct; the microscopy records alone are not treated as binding experiments. ERBB2 is an established signaling receptor, so GO:0005102 captures the partner class. EFNB1 is not an ERBB2 ligand (the reported association is not ligand-induced receptor activation), so GO:0048018 is not used. This does not assign receptor agonism, kinase activity, a purified direct interface, or a uniquely mapped binding domain to EFNB1. ERBB2 extracellular and PDZ-tail deletions did not eliminate association in the tested context.

Proposed replacements: signaling receptor binding

Summary: The EFNB1–ANKRD46 yeast two-hybrid interaction supports no more informative molecular function.

Reason: The three IntAct records comprise validated yeast two-hybrid, array, and pooled-screen records for EFNB1 and Q86W74-2. Their curated reference-sequence checksums match the current official sequences; Q86W74-2 is the displayed ANKRD46 product named isoform 1, not a product inferred from the numerical suffix. The source uses repeat testing and sequence confirmation, but the per-pair record does not establish orthogonal mammalian validation or a native physiological role. ANKRD46 has ankyrin repeats without an independently established narrower functional partner class in the inspected official record. Under the GO:0005515 protein-binding policy, a bare protein-binding row with no more informative supported activity is removed as uninformative; removal does not mean the reported interaction is false.

Summary: The BioPlex ERBB2 association supports signaling receptor binding.

Reason: The exact EFNB1-bait/ERBB2-prey record is from human HCT116 cells in PMID:33961781, with a C-terminal tandem tag on EFNB1. The curated AP-MS association and the source quality controls support the positive binding assertion; the independent endogenous co-immunoprecipitation study PMID:22279592 corroborates the pair. ERBB2 is a signaling receptor, warranting GO:0005102 without assuming receptor activation or a purified binary interface; EFNB1 is not an ERBB2 ligand, so GO:0048018 receptor ligand activity is not used. The screen does not independently establish every normal-tissue context.

Proposed replacements: signaling receptor binding

Summary: The BioPlex EPHB2 association records EFNB1 binding an Eph receptor as its ligand.

Reason: IntAct records human EFNB1 as tagged bait and EPHB2 as prey in HEK293T cells for PMID:33961781. EPHB2 is an Eph-family receptor, and the independently recovered extracellular-interaction study PMID:35922511 also reports this exact pair. The interaction is EFNB1 binding its cognate receptor class as a ligand, so the informative replacement is GO:0048018 receptor ligand activity, matching the replacement used for the GO:0046875 ephrin receptor binding rows in this review (receptor-specific binding terms are not used for ligands). AP-MS recovery establishes association in the assayed cell system, not its direct interface, stoichiometry, or a universal presentation requirement.

Proposed replacements: receptor ligand activity

Summary: The human extracellular EFNB1–EPHA2 pair records EFNB1 binding an Eph receptor as its ligand.

Reason: PMID:35922511 Supplementary Table 1 row 294 identifies P98172 and an EFNB1 construct sequence matching official residues 1–240, used with the source fusion tags. Both assay orientations are present in the raw primary matrix, and the exact pair is retained in the Figure1_network source data. EPHA2 (P29317) is independently identified as an Eph receptor. The source-data Figure2_network sheet row 120 assigns a source-measured SPR dissociation constant of 41,000 nM to this pair (source JS). This is a soluble recombinant ectodomain measurement, not a native membrane affinity. The author correction PMID:39433908 concerns other figure and methods details and does not revise these EFNB1 entries. A micromolar ectodomain affinity does not by itself show that EPHA2 is a cognate receptor that EFNB1 activates. The replacement is GO:0048018 receptor ligand activity because the term is not receptor-specific and EFNB1's ligand activity is established by its high-affinity EPHB2 binding (0.66 nM in the same dataset) and ephrin-B biology; this row adds the measured interaction only. No cross-class receptor agonism or intracellular-fragment activity is inferred.

Proposed replacements: receptor ligand activity

Summary: The human extracellular EFNB1–EPHB2 pair records EFNB1 binding an Eph receptor as its ligand.

Reason: PMID:35922511 Supplementary Table 1 row 294 identifies P98172 and an EFNB1 construct sequence matching official residues 1–240, used with the source fusion tags. Both assay orientations are present in the raw primary matrix, and the exact pair is retained in the Figure1_network source data. EPHB2 (P29323) is independently identified as an Eph receptor. The source-data Figure2_network affinity of 0.66 nM is explicitly marked Published and is not counted as a new measurement by this study. The author correction PMID:39433908 concerns other figure and methods details and does not revise these EFNB1 entries. The pair is a ligand–cognate receptor interaction, so the replacement is GO:0048018 receptor ligand activity, consistent with the GO:0046875 rows in this review; no cross-class receptor agonism or intracellular-fragment activity is inferred.

Proposed replacements: receptor ligand activity

Summary: The human extracellular EFNB1–EPHA4 pair records EFNB1 binding an Eph receptor as its ligand.

Reason: PMID:35922511 Supplementary Table 1 row 294 identifies P98172 and an EFNB1 construct sequence matching official residues 1–240, used with the source fusion tags. Both assay orientations are present in the raw primary matrix, and the exact pair is retained in the Figure1_network source data. EPHA4 (P54764) is independently identified as an Eph receptor. The source-data Figure2_network sheet row 116 assigns a source-measured SPR dissociation constant of 25,000 nM to this pair (source JS). This is a soluble recombinant ectodomain measurement, not a native membrane affinity. The author correction PMID:39433908 concerns other figure and methods details and does not revise these EFNB1 entries. A micromolar ectodomain affinity does not by itself show that EPHA4 is a cognate receptor that EFNB1 activates. The replacement is GO:0048018 receptor ligand activity because the term is not receptor-specific and EFNB1's ligand activity is established by its high-affinity EPHB2 binding (0.66 nM in the same dataset) and ephrin-B biology; this row adds the measured interaction only. No cross-class receptor agonism or intracellular-fragment activity is inferred.

Proposed replacements: receptor ligand activity

Summary: The U2OS EFNB1–EPHB2 association records EFNB1 binding an Eph receptor as its ligand.

Reason: The original U2OS AP-MS network for PMID:40205054 contains edge 587044 from EFNB1 node 553078 to EPHB2 node 575622; node metadata identify EFNB1 as bait. This is one selected pair among 39 incident EFNB1 edges, not an inference from a modeled cellular component. The study used C-terminally tagged human ORFeome constructs, anti-HA purification, protein-level and interaction filtering. EPHB2 is EFNB1's cognate Eph receptor, so the informative replacement is GO:0048018 receptor ligand activity (consistent with the GO:0046875 rows in this review), while retaining the AP-MS association and overexpression limits.

Proposed replacements: receptor ligand activity

Summary: The EFNB1–PICK1 interaction is binding of the ephrin C-terminal PDZ recognition motif by a PDZ protein.

Reason: The literal PMID:9883737 IntAct record identifies human EFNB1 and mouse PICK1 Q62083 in a positive yeast two-hybrid experiment; both reference-sequence checksums match the inspected official sequences. It is not a GRIP1 record. The abstract independently reports Eph/ephrin interactions with PDZ proteins, and PICK1 contains a PDZ domain. However, the full original could not be recovered and this exact record has no assayed-region feature, so the specific PICK1 interface is not claimed to have been independently read. The paper's own conclusion is that ephrin ligands carry PDZ recognition motifs bound by PDZ proteins, so GO:0030165 PDZ domain binding is the informative replacement for bare protein binding; no PICK1-dependent trafficking activity is inferred for EFNB1.

Proposed replacements: PDZ domain binding

Summary: Nuclear localization concerns an intracellular EFNB1 fragment.

Reason: PMID:16930449 uses a sequenced human EFNB1 cDNA in COS cells and engineered intracellular fragments. The ICD construct beginning at Val260 localizes to nucleus and cytoplasm; deletion of its N-terminal region changes that distribution. The cleavage positions used to design constructs were inferred, not established by sequencing an endogenous human cleavage product. Preserve the positive nuclear observation as a processed-fragment context without relocating the full transmembrane ligand or assigning transcription-factor activity from localization.

Summary: The inferred nuclear pool is consistent with processed intracellular-fragment evidence.

Reason: The combined IEA source includes mouse Efnb1 P52795, its Ensembl product, and the UniProt nuclear-location mapping. Current mouse experimental annotation points to PMID:19515908, whose tagged ephrin-B1 cytoplasmic domain enters the nucleus in COS-7 cells and is recruited there by ZHX2. Mouse cortical pull-downs use an ephrin-B antibody that does not uniquely distinguish B1/B2 fragments. Independent human-cDNA fragment experiments in PMID:16930449 support the same limited location. Keep this as non-core processed-fragment localization, not a claim that the intact receptor ligand is a nuclear protein.

Summary: Cytoplasmic localization is defensible for intracellular EFNB1 fragments.

Reason: The Ensembl-Compara transfer names mouse Efnb1 P52795 and its product. The current donor has a direct cytoplasmic annotation from PMID:19515908; its GFP-tagged cytoplasmic domain distributes between cytoplasm and nucleus. The independently inspected human-cDNA processing study PMID:16930449 likewise shows a cytoplasmic ICD pool. Preserve this contextual assertion without treating full-length EFNB1 as a diffuse cytosolic protein or inventing a donor error.

Summary: Retain the experimentally curated plasma-membrane location.

Reason: PMID:10197531 explicitly reports ephrin-B1 in membrane microdomains and receptor-induced raft patching. The complete original was not recovered, so detailed construct and fractionation claims are not added. The membrane-bound human protein, direct human-cell surface observations, and supported HPA plasma-membrane signal independently establish this location. There is no basis to reject the curator’s experimental annotation because access is incomplete.

Summary: The membrane-associated precursor and remnant support plasma-membrane localization.

Reason: PMID:16930449 examines human-cDNA EFNB1, its membrane-bound C-terminal remnant, and a released ICD in separate constructs and cell fractions. The source supports the membrane context from which proteolysis occurs; the nuclear/cytoplasmic ICD is a different product state. Native human-cell surface detection in PMID:12970314 independently supports plasma membrane as a principal location.

Summary: Retain the membrane-anchored ephrin ligand location.

Reason: The independently assessed complete abstract of PMID:7973638 supports membrane attachment or clustering as an important presentation context for Eph ligands. Full operative Methods were inaccessible, so no unseen construct, species, or presentation details are asserted. Human EFNB1 plasma-membrane localization is independently established by target-specific observations. The separate PMID:8798744 abstract also reports signaling with a soluble LERK2 format, so this annotation does not imply that every soluble ephrin preparation is inactive.

Summary: The inherited plasma-membrane location is conserved and directly supported.

Reason: The literal IBD at PTN000885159 carries GO:0005886, and the human P98172 leaf PTN002477887 descends from this node in the recovered 389-node PTHR11304 tree. Its ungapped 346-residue alignment sequence matches the official sequence exactly. Experimental target and descendant evidence grounds the placement; inclusion of P98172 among IBD sources is valid target grounding, not circularity. Direct human-cell evidence independently corroborates plasma membrane.

Summary: HPA supports plasma-membrane localization of human EFNB1.

Reason: The target-specific HPA record for ENSG00000090776/P98172 reports plasma membrane with per-location reliability supported, using HPA067188 in HaCaT, PC-3 and U2OS cells. The XML and target HTML were inspected; the inference is based on the curated antibody-assay record rather than an independent pixel interpretation. This is positive corroboration of the membrane ligand location.

Summary: The UniProt plasma-membrane mapping is consistent with direct evidence.

Reason: The electronic mapping from UniProt subcellular location SL-0039 is supported by the official transmembrane topology and human-cell surface observations. The soluble intracellular processing products are separately described and do not invalidate localization of the mature membrane-bound ligand.

Summary: The historical LERK2 description supports a type-I membrane ligand.

Reason: Although the title of PMID:8660976 foregrounds other LERK genes, its abstract explicitly discusses LERK2/EPLG2 as a transmembrane ligand. This is not evidence of a wrong-gene attribution. Preserve the traceable membrane statement, corroborated by the current EFNB1 identity and independent human-cell localization; no uninspected experimental detail is inferred.

Summary: EFNB1 is a plasma-membrane participant in the curated event EPH receptors autophosphorylate.

Reason: The saved Reactome:R-HSA-3928578 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928578 → 3928372 → 9028356 → 3928331 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event EPH:EFN tetramers oligomerise.

Reason: The saved Reactome:R-HSA-3928591 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928591 → 3928376 → 3928331 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event p190RhoGEF exchanges GTP for GDP on RHOA, activating it.

Reason: The saved Reactome:R-HSA-3928592 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928592 → 3928573 → 3928563 → 3928549 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event N-WASP binds ARP2/3 and G-actin.

Reason: The saved Reactome:R-HSA-3928595 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928595 → 3928536 → 3928566 → 3928511 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event EPH:EFN dimers tetramerise.

Reason: The saved Reactome:R-HSA-3928597 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928597 → 3928376 → 3928331 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event Recruitment of p190RhoGEF to p-FAK.

Reason: The saved Reactome:R-HSA-3928598 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928598 → 3928549 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event CDC42 and PIP2 bind WASL, activating it.

Reason: The saved Reactome:R-HSA-3928601 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928601 → 3928566 → 3928511 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event SFKs phosphorylate VAV2,3.

Reason: The saved Reactome:R-HSA-3928604 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928604 → 3928540 → 3928499 → 3928351 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event p-EPHs bind VAV2,3.

Reason: The saved Reactome:R-HSA-3928607 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928607 → 3928499 → 3928351 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event KALRN exchanges GTP for GDP on RAC1, activating it.

Reason: The saved Reactome:R-HSA-3928612 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928612 → 3928586 → 3928533 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event EFNB:GRB4 binds GIT1.

Reason: The saved Reactome:R-HSA-3928614 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928614 → 3928522 → 3928510 → 3928571 → 3928495 → 3928412. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event EFNBs bind SFKs.

Reason: The saved Reactome:R-HSA-3928615 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928615 → 3928331 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event EPHB2 binds KALRN.

Reason: The saved Reactome:R-HSA-3928617 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928617 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event EPHBs bind EFNBs.

Reason: The saved Reactome:R-HSA-3928624 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928624 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event PAKs autophosphorylate.

Reason: The saved Reactome:R-HSA-3928625 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928625 → 3928537 → 3928560 → 3928512 → 3928559 → 3928522 → 3928510 → 3928571 → 3928495 → 3928412. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event EPHB phosphorylates TIAM1.

Reason: The saved Reactome:R-HSA-3928627 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928627 → 3928541 → 3928531 → 3928492 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event VAV exchanges GTP for GDP on RAC1, activating it.

Reason: The saved Reactome:R-HSA-3928628 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928628 → 3928590 → 3928532 → 3928499 → 3928351 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event GIT1 binds bPIX.

Reason: The saved Reactome:R-HSA-3928631 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928631 → 3928559 → 3928522 → 3928510 → 3928571 → 3928495 → 3928412. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event bPIX exchanges GTP for GDP on RAC, activating it.

Reason: The saved Reactome:R-HSA-3928633 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928633 → 3928512 → 3928559 → 3928522 → 3928510 → 3928571 → 3928495 → 3928412. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event p-EFNB binds GRB4.

Reason: The saved Reactome:R-HSA-3928639 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928639 → 3928510 → 3928571 → 3928495 → 3928412. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event PAKs phosphorylate MLC.

Reason: The saved Reactome:R-HSA-3928640 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928640 → 3928609 → 3928534 → 3928560 → 3928512 → 3928559 → 3928522 → 3928510 → 3928571 → 3928495 → 3928412. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event PAK binds RAC and bPIX.

Reason: The saved Reactome:R-HSA-3928641 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928641 → 3928560 → 3928512 → 3928559 → 3928522 → 3928510 → 3928571 → 3928495 → 3928412. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event EPHB:NMDAR binds TIAM1.

Reason: The saved Reactome:R-HSA-3928645 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928645 → 3928531 → 3928492 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event Clathrin internalises EPH:EFN complexes.

Reason: The saved Reactome:R-HSA-3928654 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928654 → 3928331 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event MMP2,9 cleave EPHB.

Reason: The saved Reactome:R-HSA-3928657 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 3928657 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event EFNBs bind SDCBP.

Reason: The saved Reactome:R-HSA-4093329 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 4093329 → 3928331 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event Ras:GTP binds p120-RasGAP.

Reason: The saved Reactome:R-HSA-4093331 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 4093331 → 4093322 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event p-EPHB phosphorylates SDC2.

Reason: The saved Reactome:R-HSA-4093332 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 4093332 → 4093337 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event p-TIAM1 exchanges GTP for GDP on RAC1, activating it.

Reason: The saved Reactome:R-HSA-4093336 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 4093336 → 4093333 → 3928518 → 3928531 → 3928492 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: EFNB1 is a plasma-membrane participant in the curated event p120-RasGAP activates GTP hydrolysis on RAS, inactivating it.

Reason: The saved Reactome:R-HSA-4093339 participant graph resolves through component, member or polymer-unit links to human EFNB1 P98172, with plasma membrane explicitly attached to that target entity. The shortest recorded path is 4093339 → 4093319 → 4093322 → 9028348 → 3928369 → 3928378 → 2682327. This preserves the source-specific location assertion. Catalytic reactions or downstream proteins named in the event are not transferred as EFNB1 molecular activities, and membership in an EFNB set is not treated as a new target-specific biochemical experiment.

Summary: Membrane EFNB1–EPHB3 engagement supports the specific cell-cell adhesion process.

Reason: The complete abstract of PMID:8798744 reports aggregates when HEK2-expressing cells are coincubated with LERK-2-expressing cells, with inhibition by soluble ligand. Official identities resolve LERK-2 to EFNB1 and HEK2 to EPHB3/P54753. GO:0098609 describes attachment between cells more precisely than GO:0007155, which also covers attachment to extracellular substrates. EFNB1 supplies the receptor-binding ligand at the adhesive contact, so this is direct participation rather than an inference from necessity alone. Full Methods were not recovered; the retained TAS assertion does not establish universal adhesion rather than repulsion, native tissue behavior, or unseen construct and soluble-ligand formulation details.

Proposed replacements: cell-cell adhesion

Summary: The specific ephrin receptor pathway describes the demonstrated intercellular signaling.

Reason: PMID:8798744 directly links LERK2/EFNB1 binding to HEK2/EPHB3 with receptor tyrosine phosphorylation and contact-dependent behavior. GO:0048013 describes the series of signals initiated by ephrin–Eph engagement more informatively than generic cell–cell signaling. This is a source-based biological refinement, not a claim that the two terms have a direct is_a relationship. The reported soluble-ligand activity is retained with unresolved formulation details because full Methods were inaccessible.

Proposed replacements: ephrin receptor signaling pathway

Summary: Retain the conserved axon-guidance role as a developmental context.

Reason: The literal GO:0007411 IBD at PTN000885159 is ancestral to the exact human sequence and has multiple experimental descendant sources, including mouse Efnb1 MGI:102708. The current mouse record independently retains IMP axon-guidance evidence. This is a curated ancestral functional placement, not pairwise transfer judged by donor count. There is no target-specific loss evidence. Keep the biologically defensible neural-development context as non-core relative to the general ephrin receptor-binding and signaling mechanism; no human axon assay is invented.

Summary: Native human granulosa cells display ephrin-B1 at the cell surface.

Reason: PMID:12970314 reports flow-cytometric detection of ephrin-B1 on luteinizing granulosa cells from human patients and binding of an EphB2-Fc reagent. Selected original Methods identify an antibody raised against mouse extracellular ephrin-B1 that cross-reacts with human and rat, and a recombinant mouse EphB2–human Fc reagent. These reagent species do not change the human origin of the observed cells. The experimental surface assertion is retained without assigning a new ovarian developmental process.

Summary: Plasma membrane is a supported refinement of the broad membrane annotation.

Reason: IPR001799 and IPR019765 provide ephrin receptor-binding-domain and conserved-site mappings to membrane. Those domain mappings alone do not specify every intracellular product location. Target-specific human-cell surface detection, HPA plasma-membrane support, and the mature transmembrane ligand establish GO:0005886 as a more useful location for this source assertion. Separate ICD nucleus/cytoplasm rows remain intact, so the refinement does not erase processed-product pools.

Proposed replacements: plasma membrane

Summary: Presynaptic membrane is a supported neuronal context of ephrin-B signaling.

Reason: The literal GO:0042734 IBD is at PTN002601663, on the human target path, with rat Efnb1 RGD:2540 as an experimental descendant. The current donor record cites PMID:19915143; independently recovered original Results show ephrin-B1 colocalization with VGlut1-positive presynaptic puncta, and main Methods identify E17–E18 rat cortical cultures. The current donor extension names hippocampus, a distinction retained rather than silently equated to the inspected cortical culture. A single well-grounded donor is not a weakness of the PAINT inference. Keep the neuronal specialization without claiming a directly measured human presynaptic pool.

Summary: Retain receptor-associated raft localization as a membrane subdomain context.

Reason: The complete abstract of PMID:10197531 directly reports ephrin-B1 localization to raft membrane microdomains, GRIP-family recruitment and EphB2-induced raft patching. The full operative body remains inaccessible, so exact organism/construct/fractionation details are not invented. Preserve this positive experimental assertion as a specialized membrane context, corroborated by the mouse donor annotation, rather than reject it from incomplete access.

Summary: The electronic raft annotation has positive ortholog and location-mapping support.

Reason: The combined source includes mouse Efnb1 P52795 and the SL-0370 raft mapping. Current mouse experimental GO:0045121 cites PMID:10197531, whose abstract explicitly describes ephrin-B1 raft localization. This agrees with the target experimental row and supports the transfer. The non-core designation concerns a particular membrane organization, not an error in the donor or an assumption that all EFNB1 is raft-restricted.

Summary: Retain the ortholog-supported synaptic location as a neural context.

Reason: The manually transferred ISS source is rat Efnb1 P52796, whose current GO:0045202 IDA cites PMID:9883737. The cited abstract reports Eph/ephrin and PDZ-protein synaptic colocalization; independent rat cortical data in PMID:19915143 support ephrin-B1 at synaptic specializations. Full PMID:9883737 Methods were not recovered, so the original annotation is not narrowed to a particular side of the synapse or assigned to a measured human neuronal product.

Summary: The ancestral ephrin receptor-binding annotation is refined to receptor ligand activity for human EFNB1.

Reason: The recovered tree and MSA place the exact human 346-residue sequence below the GO:0046875 IBD at PTN000885159. The literal IBD includes experimentally grounded descendants and the target itself. Target inclusion is appropriate, not circular. Because ephrin-B1 functions as a signaling ligand that activates Eph receptor tyrosine kinases rather than a passive binding partner, GO:0048018 (receptor ligand activity) captures its primary molecular function more accurately than the receptor-specific binding term GO:0046875.

Proposed replacements: receptor ligand activity

Summary: The human MSC–T-cell study supports EFNB1 acting as an Eph receptor ligand.

Reason: PMID:23711177 identifies ephrin-B1 on human T cells and EphB2 on human mesenchymal stromal cells, with receptor-Fc engagement and blocking experiments in the recovered Results. The normal cache is abstract-only and selected original text, rather than the complete Methods or all figures, was accessed independently. Because ephrin-B1 acts as a signaling ligand engaging Eph receptors rather than a passive binding partner, GO:0048018 (receptor ligand activity) captures its primary molecular function more accurately than the receptor-specific binding term GO:0046875, without importing every immunosuppressive downstream effect or claiming a purified full-length pair.

Proposed replacements: receptor ligand activity

Summary: Ortholog and ephrin-B ectodomain evidence support receptor ligand activity.

Reason: The combined IEA cites mouse Efnb1 P52795 and IPR034255. Current InterPro explicitly maps the ephrin-B ectodomain to GO:0046875, and the mouse record has experimental Eph receptor interactions. Because ephrin-B1 functions through its extracellular domain as a signaling ligand that activates Eph receptors, GO:0048018 (receptor ligand activity) represents its primary molecular function more accurately than receptor-specific binding GO:0046875. The inference concerns the extracellular ligand domain, not the released intracellular fragment.

Proposed replacements: receptor ligand activity

Summary: LERK2 binding and activation of HEK2 supports receptor ligand activity for EFNB1.

Reason: The independently assessed complete PMID:8798744 abstract directly reports LERK2 binding and signaling with HEK2, promoting receptor tyrosine kinase activity. Official current identities resolve these names to EFNB1/P98172 and EPHB3/P54753. Because EFNB1 acts as the signaling ligand that activates the Eph receptor rather than merely binding it, GO:0048018 (receptor ligand activity) captures the primary molecular function more accurately than GO:0046875. Full Methods were unavailable, so the soluble-versus-membrane comparison is not recast as a universal oligomerization rule or assigned unseen construct details.

Proposed replacements: receptor ligand activity

Summary: Human EFNB1 inherits the established ephrin receptor signaling function.

Reason: The literal GO:0048013 IBD at PTN000885159 lies on the exact human sequence path and is experimentally grounded in descendant ephrin proteins. This is an ancestral PAINT assertion, not a numerical donor-vote or similarity shortcut. The target directly binds Eph receptors, and source-specific signaling evidence corroborates participation. It does not confer the receptor kinase activity on EFNB1.

Summary: The ephrin-B ectodomain mapping is consistent with pathway participation.

Reason: The current IPR034255 record explicitly maps to GO:0048013. Human EFNB1 provides the extracellular Eph-binding ligand, a step of the signaling mechanism itself, rather than merely being required for an unrelated outcome. Independent human receptor interaction evidence supports this mapping. The annotation does not imply that every cleaved EFNB1 fragment retains the extracellular signaling function.

Summary: Human urinary exosome proteomics detected ephrin-B1.

Reason: The original laboratory database supporting PMID:19056867 lists EFNB1, ephrin-B1 precursor, RefSeq NP_004420, and one peptide. Its final value 2 is the bibliography reference number for the 2009 study, not a second peptide or replicate count. The study surveys a low-density urinary membrane fraction chiefly composed of exosomes from healthy humans. Preserve the positive HDA detection as a contextual extracellular-vesicle location; it does not establish an active exosomal signaling function, a particular processing chain, or independent peptide-level revalidation.

Summary: The inherited glutamatergic-synapse location is supported in the neuronal context.

Reason: The literal GO:0098978 IBD at PTN002601663 is ancestral to human EFNB1 and is grounded in rat Efnb1 RGD:2540. Current donor annotations and the independently recovered PMID:19915143 Results document EFNB1 at VGlut1-positive excitatory synaptic puncta. Main Methods specify embryonic rat cortical neurons; the donor annotation’s hippocampal extension is not silently substituted for that source scope. Retain the valid phylogenetic location as a neuronal specialization without inventing a human localization assay or treating the single donor as weak evidence.