{"citation":"GlycoDisease: a source-auditable resource for candidate protein\u2013disease relationships and quantitative glycosylation evidence","contact":"zhangyang@fudan.edu.cn","data_license":"Pending author confirmation","data_release_id":"v2.0.0-rc.3.identity.5","data_sources":["PubMed literature (LLM extraction)","UniProt database","GlyGen database","GlyTouCan database"],"database":"GlycoDisease Candidate Relationship Resource","disease_categories":["cancer","neurodegenerative","cardiovascular","metabolic","autoimmune","infection","liver","kidney"],"interpretation_version":"v2.0.0-rc.3.context.4","joint_release":{"author_declarations_confirmed":false,"coordinate_version":"source-table-audit.9","data_release_id":"v2.0.0-rc.3.identity.5","historical_model_not_retrained_on_current_data":true,"independent_biological_accuracy_established":false,"interpretation_version":"v2.0.0-rc.3.context.4","joint_release_id":"joint-2026-09-06.1","knowledge_base_sha256":"73aaf23c7b9823a596cbcbcb274e72bf30c8dffa17d515cb966b843ae08736d8","manuscript_build_report_sha256":"0641cbe28a56722836627684ab88491484840efb6051cd3523b9329e16b7d44d","manuscript_files":[{"bytes":899126,"path":"GlycoDisease_joint_manuscript.docx","sha256":"f6895bfce596bbfbd086268b8f1aba7727cebc8c0da566b603c667f9d16f9c23"},{"bytes":46183,"path":"GlycoDisease_joint_supplement.docx","sha256":"4a431ab7e5391a315efbc730e399f694e2c864f8c719a5196fd05eae7735757f"},{"bytes":800586,"path":"GlycoDisease_joint_manuscript.pdf","sha256":"6c2d096d1e7e4c265054ee9b24bc0f83ad8281fec788389e4adbd0d9d3e89f39"},{"bytes":264649,"path":"GlycoDisease_joint_supplement.pdf","sha256":"7b94e87c50b5160b99019352d80ddf3c67449c4b08ccf04251d58fc7e2595dd9"},{"bytes":12342073,"path":"GlycoDisease_Supplementary_Data_S1-S8.zip","sha256":"0562dae060f66e477eb9483107dcb2d55cd7c94e8d72cb312cc4696199aed99e"}],"manuscript_submitted":false,"model_release_id":"v2.0.0-rc.1","model_sha256":"80c1ee1d6e7ecfed1587bf2d237a51efc93fc63410d417304d13625310fe9f99","permanent_archive_doi":null,"platform_release_id":"web-2026-09-06.1","public_deployment_verification":"recorded_separately_after_staging","public_url":"https://glycodb.aicopilot.cool/","repository_visibility":"private_by_author_choice","runtime_manifest_sha256":"31efa564634b0fc9a998bad0a892d9cfd274337fda138bc6522a87318f59b491","schema_version":"2026-09-06.joint-release.1","source_commit":"2ef93e5c6ad1b5dd6e95c06f8c741178720c536a","source_coordinate_and_exclusion_questions_open":true,"source_event_version":"source_event_core10","source_repository":"https://github.com/HangZhouSheep/GD","statistics_version":"source_statistics11","status":"BOUND_AUTHOR_REVIEW_RELEASE","title":"GlycoDisease: a source-auditable resource for candidate protein\u2013disease relationships and quantitative glycosylation evidence"},"joint_release_id":"joint-2026-09-06.1","knowledge_base_sha256":"73aaf23c7b9823a596cbcbcb274e72bf30c8dffa17d515cb966b843ae08736d8","manuscript_scripts_commit":"2ef93e5c6ad1b5dd6e95c06f8c741178720c536a","manuscript_scripts_license":"MIT","model_contract":{"aggregate_metrics":{"macro_auprc":0.1645127617173775,"macro_auroc":0.713942411873467,"macro_observed_positive_fraction":0.0461174405160548},"category_order":["cancer","neurodegenerative","cardiovascular","metabolic","autoimmune","infection","liver","kidney"],"current_data_performance_measured":false,"evaluation_design":"v2 literature-derived candidate-P versus archived unlabeled ranking on a reused CD-HIT 40% cluster holdout","historical_model_not_retrained_on_current_data":true,"holdout_novelty":"reused_historical_split_not_newly_sealed","model_sha256":"80c1ee1d6e7ecfed1587bf2d237a51efc93fc63410d417304d13625310fe9f99","model_version":"glycodisease-v2-pu-ranker-2026-08-30.1","representation_version":"esm2_t12_35M_UR50D:first-1022-aa:layer-12:mean-pool","score_semantics":"Category-specific literature-derived candidate-P-versus-unlabeled ranking score; not a probability.","utility_summary":{"categories_auprc_above_random_baseline":8,"categories_auprc_ci_low_above_random_baseline":7,"categories_auroc_ci_low_above_chance":7,"category_count":8,"interpretation":"Moderate, heterogeneous candidate-ranking signal relative to random ranking; not uniformly strong classification.","macro_auprc_random_baseline":0.0461174405160548,"macro_auprc_to_baseline_ratio":3.5672569829651737,"top_decile_enrichment_max":4.59030303030303,"top_decile_enrichment_min":1.7576470588235293},"validated_sequence_length":[10,1022]},"model_is_historical_not_retrained_on_current_data":true,"model_release_id":"v2.0.0-rc.1","model_sha256":"80c1ee1d6e7ecfed1587bf2d237a51efc93fc63410d417304d13625310fe9f99","platform_code_license":"MIT","platform_release_id":"web-2026-09-06.1","platform_source_commit":"2ef93e5c6ad1b5dd6e95c06f8c741178720c536a","quality_contract":{"approximate_design_95_interval":[0.9582028118022285,1.0],"engineering_tests_in_scientific_handoff":439,"independent_accuracy_measured":false,"mandatory_human_adjudication":false,"quality_gate_passed":false,"recommended_term":"same-model AI evidence-support consistency","release_gate_note":"Historical AI technical-failure summary retained; mandatory human adjudication is not a release prerequisite.","sample_qc_status":"AI_ONLY_SAMPLE_QC_COMPLETE_WITH_TECHNICAL_FAILURES","sample_rows":200,"source_identity_supported_scope_rows":48,"stratified_weighted_support_estimate":0.9791418151767748,"strictly_supported_rows":193,"technical_failures_counted_as_unsupported":true,"unweighted_support_proportion":0.965},"release_date":"2026-09-06","release_id":"web-2026-09-06.1","runtime_manifest_sha256":"31efa564634b0fc9a998bad0a892d9cfd274337fda138bc6522a87318f59b491","source_commit":"2ef93e5c6ad1b5dd6e95c06f8c741178720c536a","source_event_summary":{"BH_sensitivity_filter_rows_all":739,"BH_sensitivity_filter_rows_consistent_subset":729,"all_reported_log2fc_reproduced_from_displayed9v9_channels":true,"all_site_labels_preserved":true,"coordinate_status_counts":{"multiple_exact_peptide_locations_not_unique_coordinate":7,"peptide_not_in_archived_canonical_sequence":1,"reported_site_matches_single_literal_marker_not_biological_validation":2670,"reported_site_not_among_literal_marked_residue_coordinates":46},"do_not_add_table_rows_to_original_extraction_count":true,"independent_accuracy_or_biological_novelty":false,"new_source_table_layer_rows":2724,"new_study_count":1,"original_assertion_count_stays":31808,"other_rows_retained_with_warnings":59,"raw_MS_or_protein_normalization_recomputed":false,"same_study_as_previously_inspected_AD_case":true,"sample_exclusion_question_unresolved":true,"sequence_and_name_consistent_rows":2665,"source_accessions":1037,"source_direction_counts":{"decrease":2106,"increase":618},"source_pmcid":"PMC12529471","threshold_note":"|log2FC|>0.5, raw p<0.05; BH q computed only as a secondary all2724-row sensitivity analysis, not substituted for publication.","uncorrected_filter_rows_all":743,"uncorrected_filter_rows_consistent_subset":733},"source_event_version":"source_event_core10","statistics":{"articles_contributing_records":2813,"articles_processed":4273,"explicit_glycan_support_records":750,"glycan_structures":1960,"glycan_support_status":{"no":31056,"uncertain":2,"yes":750},"total_disease_strings":7566,"total_proteins":9057,"total_relationships":31808},"version":"2026-09-06","website_profile":{"coverage":{"ambiguous_or_unmapped_protein_rows":{"count":23083,"denominator":31808,"proportion":0.7256979376257545},"evidence_row_protein_mapping":{"denominator":31808,"mapped":8725,"proportion":0.2743020623742455},"other_or_unmapped_disease_rows":{"count":17570,"denominator":31808,"proportion":0.5523767605633803},"protein_record_archived_accession":{"denominator":9057,"mapped_unique_uniprot_ids":1624,"proportion":0.1793088219057083}},"data_release_id":"v2.0.0-rc.3.identity.5","model_release_id":"v2.0.0-rc.1","platform_release_id":"web-2026-09-06.1","record_counts":{"articles_contributing_records":2813,"articles_entering_extraction":4273,"candidate_evidence_records":31808,"disease_strings":7566,"explicit_glycan_support_articles":246,"glycan_support_status":{"no":31056,"uncertain":2,"yes":750},"linked_glycan_structures":1960,"protein_records":9057},"record_semantics":{"all_records":"AI-extracted candidate protein-disease evidence records from a glycoscience-focused corpus","explicit_glycan_support":"Rows whose supplied source context explicitly supports glycan or glycosylation involvement","glycan_support_is_database_inclusion_requirement":false,"independent_human_gold_standard_complete":false,"whole_article_recall_measured":false},"resource_name":"GlycoDisease Candidate Relationship Resource","resource_scope":"Glycoscience-corpus-derived candidate protein-disease evidence with explicit glycan-involvement stratification","schema_version":"2026-09-02.glycodisease-website-profile.1"}}