Integrating machine learning algorithms and multiple immunohistochemistry validation to unveil novel diagnostic markers based on costimulatory molecules for predicting immune microenvironment status in triple-negative breast cancer.

{"sci_version": "1.2", "schema": "https://corpus.bntrg.co.uk/sci/1.0/sci.schema.json", "id": "pmid:39007147", "kind": "science-record", "revision": 4, "content_hash": "sha256:8b095d4e2138b54ef3fdb46ae8309d7291a2bab1353ab22222aee16dc2d0311b", "created_at": "2026-11-01T20:29:05Z", "updated_at": "2026-11-01T22:54:42Z", "generator": {"name": "corpus-assay-engine", "run_id": "reading-sci-2026-11-01"}, "paper": {"state": "complete", "title": "Integrating machine learning algorithms and multiple immunohistochemistry validation to unveil novel diagnostic markers based on costimulatory molecules for predicting immune microenvironment status in triple-negative breast cancer.", "doi": "10.3389/fimmu.2024.1424259", "pmid": "39007147", "authors": ["Chao Zhang", "Wenyu Zhai", "Yuyu Ma", "Minqing Wu", "Qiaoting Cai", "Jiajia Huang", "Zhihuan Zhou", "Fangfang Duan"], "journal": "Frontiers in immunology", "year": 2024, "abstract": "Introduction: Costimulatory molecules are putative novel targets or potential additions to current available immunotherapy, but their expression patterns and clinical value in triple-negative breast cancer (TNBC) are to be clarified.\nMethods: The gene expression profiles datasets of TNBC patients were obtained from The Cancer Genome Atlas and the Gene Expression Omnibus databases. Diagnostic biomarkers for stratifying individualized tumor immune microenvironment (TIME) were identified using the Least Absolute Shrinkage and Selection Operator (LASSO) and Support Vector Machine-Recursive Feature Elimination (SVM-RFE) algorithms. Additionally, we explored their associations with response to immunotherapy via the multiplex immunohistochemistry (mIHC).\nResults: A total of 60 costimulatory molecule genes (CMGs) were obtained, and we determined two different TIME subclasses (\"hot\" and \"cold\") through the K-means clustering method. The \"hot\" tumors presented a higher infiltration of activated immune cells, i.e., CD4 memory-activated T cells, resting NK cells, M1 macrophages, and CD8 T cells, thereby enriched in the B cell and T cell receptor signaling pathways. LASSO and SVM-RFE algorithms identified three CMGs (CD86, TNFRSF17 and TNFRSF1B) as diagnostic biomarkers. Following, a novel diagnostic nomogram was constructed for predicting individualized TIME status and was validated with good predictive accuracy in TCGA, GSE76250 and GSE58812 databases. Further mIHC conformed that TNBC patients with high CD86, TNFRSF17 and TNFRSF1B levels tended to respond to immunotherapy.\nConclusion: This study supplemented evidence about the value of CMGs in TNBC. In addition, CD86, TNFRSF17 and TNFRSF1B were found as potential biomarkers, significantly promoting TNBC patient selection for immunotherapeutic guidance.", "claims": [{"id": "claim-method", "statement": "In the TCGA and GSE76250, to avoid possible influence of multicollinearity, we firstly conducted the LASSO logistic regression analysis through the “glmnet” package to screen out biomarkers from above all 60 CMGs at the optimal value of log lambda with the smallest classification error (30). Besides, the SVM-RFE machine learning algorithm based on the support vector machine was utilized to identify the most valuable biomarkers from all 60 CMGs by subtracting the feature vector determined using SVM with the “e1071” and “caret” R packages (31).", "kind": "methodological", "target": null, "evidence": [{"asset": "source:pmid:39007147", "locator": "sections[7] \"Screening and identification of the diagnostic CMGs biomarkers\" paragraph 2", "quote": "In the TCGA and GSE76250, to avoid possible influence of multicollinearity, we firstly conducted the LASSO logistic regression analysis through the “glmnet” package to screen out biomarkers from above all 60 CMGs at the optimal value of log lambda with the smallest classification error (30). Besides, the SVM-RFE machine learning algorithm based on the support vector machine was utilized to identify the most valuable biomarkers from all 60 CMGs by subtracting the feature vector determined using SVM with the “e1071” and “caret” R packages (31)."}], "state": "complete"}, {"id": "claim-data-access", "statement": "The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material .", "kind": "methodological", "target": null, "evidence": [{"asset": "source:pmid:39007147", "locator": "sections[21] \"Data availability statement\" paragraph 2", "quote": "The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material ."}], "state": "complete"}, {"id": "claim-target-result", "statement": "LASSO and SVM-RFE algorithms identified three CMGs (CD86, TNFRSF17 and TNFRSF1B) as diagnostic biomarkers.", "kind": "qualitative", "target": null, "evidence": [{"asset": "source:pmid:39007147", "locator": "sections[0] \"Abstract\" paragraph 3", "quote": "LASSO and SVM-RFE algorithms identified three CMGs (CD86, TNFRSF17 and TNFRSF1B) as diagnostic biomarkers."}], "state": "complete"}], "gaps": []}, "experiment": {"state": "complete", "objective": "LASSO and SVM-RFE algorithms identified three CMGs (CD86, TNFRSF17 and TNFRSF1B) as diagnostic biomarkers.", "steps": [{"id": "read_published_estimate", "order": 1, "name": "Read the published estimate from the source", "kind": "other", "state": "complete", "blocked_on": null, "description": "The claim, method and evidence span published with the source article.", "duration_hours": null, "parameters": {}, "inputs": ["source:pmid:39007147"], "outputs": ["the published estimate and its evidence"], "instruments": [], "reagents": []}, {"id": "obtain_source_data", "order": 2, "name": "Obtain the source data from the named repository", "kind": "data-preparation", "state": "complete", "blocked_on": null, "description": "The GEO series accession resolves and its sample accessions are listed. This lane did not download the dataset or read its licence.", "duration_hours": null, "parameters": {}, "inputs": ["data:geo:GSE58812"], "outputs": ["the source data with its repository provenance"], "instruments": [], "reagents": []}, {"id": "run_analysis", "order": 3, "name": "Run the published analysis", "kind": "analysis", "state": "complete", "blocked_on": null, "description": "Apply the paper's statistical method to the resolved series and compare the result with the published target.", "duration_hours": null, "parameters": {}, "inputs": ["data:geo:GSE58812"], "outputs": ["the replicated result and its comparison"], "instruments": [], "reagents": []}], "materials": [], "instruments": [], "reagents": [], "conditions": {}, "gaps": []}, "data": {"state": "partial", "assets": [{"id": "source:pmid:39007147", "role": "manuscript", "format": "xml", "media_type": "application/xml", "uri": "https://doi.org/10.3389/fimmu.2024.1424259", "sha256": "sha256:a37b0d1885e14ae895db12e6823ed7edebb050ac5992e3a48f807ec5e812e61c", "bytes": 57712, "access": "open", "state": "resolved", "reader": null, "reader_status": "reader_required", "identifier": "10.3389/fimmu.2024.1424259", "version": null, "note": "Full article fetched from PMC; the checksum covers the rendered section body, not the original XML bytes.", "structure": {"pmcid": "PMC11239375", "source": "europe_pmc_fulltext", "coverage": "whole_body", "text_hash_scope": "rendered_section_body"}}, {"id": "data:geo:GSE58812", "role": "raw", "format": "geo-series", "media_type": "application/vnd.ncbi.geo-series+json", "uri": "https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE58812", "sha256": null, "bytes": null, "access": "open", "state": "resolved", "reader": null, "reader_status": "reader_required", "identifier": "GSE58812", "version": null, "note": "The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material .", "structure": {"repository": "GEO", "access_route": "GEO accession URL", "access_statement_verbatim": "The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material .", "access_statement_kind": "data_availability_statement", "sample_accessions": ["GSM1419942", "GSM1419943", "GSM1419944", "GSM1419945", "GSM1419946", "GSM1419947", "GSM1419948", "GSM1419949", "GSM1419950", "GSM1419951", "GSM1419952", "GSM1419953", "GSM1419954", "GSM1419955", "GSM1419956", "GSM1419957", "GSM1419958", "GSM1419968", "GSM1419969", "GSM1419970", "GSM1419971", "GSM1419972", "GSM1419973", "GSM1419974", "GSM1419975", "GSM1419976", "GSM1419977", "GSM1419978", "GSM1419979", "GSM1419980", "GSM1419981", "GSM1419982", "GSM1419983", "GSM1419984", "GSM1419985", "GSM1419986", "GSM1419987", "GSM1419988", "GSM1419989", "GSM1419990", "GSM1419991", "GSM1419992", "GSM1419993", "GSM1419994", "GSM1419995", "GSM1419996", "GSM1419997", "GSM1419998", "GSM1419999", "GSM1420000", "GSM1420001", "GSM1420002", "GSM1420003", "GSM1420004", "GSM1420005", "GSM1420006", "GSM1420007", "GSM1420008", "GSM1420009", "GSM1420010", "GSM1420011", "GSM1420012", "GSM1420013", "GSM1420014", "GSM1420015", "GSM1420016", "GSM1420017", "GSM1420018", "GSM1420019", "GSM1420020", "GSM1420021", "GSM1420022", "GSM1420023", "GSM1420024", "GSM1420025", "GSM1420026", "GSM1420027", "GSM1420028", "GSM1420029", "GSM1420030", "GSM1420031", "GSM1420032", "GSM1420033", "GSM1420034", "GSM1420035", "GSM1420036", "GSM1420037", "GSM1420038", "GSM1420039", "GSM1420040", "GSM1420041", "GSM1420042", "GSM1420043", "GSM1420044", "GSM1420045", "GSM1420046", "GSM1420047", "GSM1420048", "GSM1420049", "GSM1420050", "GSM1420051", "GSM1420052", "GSM1420053", "GSM1420054", "GSM1420055", "GSM1420056", "GSM1420057"], "sample_count": 107, "group_sizes": null, "group_sizes_basis": null, "platform": "570", "taxon": "Homo sapiens", "title": "Gene-expression molecular subtyping of triple-negative breast cancer tumours: importance of immune response", "resolution_checked_at": "2026-11-01T15:39:54Z", "resolution_check": "identifier_resolution_only_no_download_no_licence_read", "downloaded": false, "licence_read": false}}, {"id": "data:geo:GSE76250", "role": "raw", "format": "geo-series", "media_type": "application/vnd.ncbi.geo-series+json", "uri": "https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE76250", "sha256": null, "bytes": null, "access": "open", "state": "resolved", "reader": null, "reader_status": "reader_required", "identifier": "GSE76250", "version": null, "note": "The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material .", "structure": {"repository": "GEO", "access_route": "GEO accession URL", "access_statement_verbatim": "The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material .", "access_statement_kind": "data_availability_statement", "sample_accessions": ["GSM1977646", "GSM1977647", "GSM1977648", "GSM1977649", "GSM1977650", "GSM1977651", "GSM1977652", "GSM1977653", "GSM1977654", "GSM1977655", "GSM1977656", "GSM1977657", "GSM1977658", "GSM1977659", "GSM1977660", "GSM1977661", "GSM1977662", "GSM1977663", "GSM1977664", "GSM1977665", "GSM1977666", "GSM1977667", "GSM1977668", "GSM1977669", "GSM1977670", "GSM1977671", "GSM1977672", "GSM1977673", "GSM1977674", "GSM1977675", "GSM1977676", "GSM1977677", "GSM1977678", "GSM1977679", "GSM1977680", "GSM1977681", "GSM1977682", "GSM1977683", "GSM1977684", "GSM1977685", "GSM1977686", "GSM1977687", "GSM1977688", "GSM1977689", "GSM1977690", "GSM1977691", "GSM1977692", "GSM1977693", "GSM1977694", "GSM1977695", "GSM1977696", "GSM1977697", "GSM1977698", "GSM1977699", "GSM1977700", "GSM1977701", "GSM1977702", "GSM1977703", "GSM1977704", "GSM1977705", "GSM1977706", "GSM1977707", "GSM1977708", "GSM1977709", "GSM1977710", "GSM1977711", "GSM1977712", "GSM1977713", "GSM1977714", "GSM1977715", "GSM1977716", "GSM1977717", "GSM1977718", "GSM1977719", "GSM1977720", "GSM1977721", "GSM1977722", "GSM1977723", "GSM1977724", "GSM1977725", "GSM1977726", "GSM1977727", "GSM1977728", "GSM1977729", "GSM1977730", "GSM1977731", "GSM1977732", "GSM1977733", "GSM1977734", "GSM1977735", "GSM1977736", "GSM1977737", "GSM1977738", "GSM1977739", "GSM1977740", "GSM1977741", "GSM1977742", "GSM1977743", "GSM1977744", "GSM1977745", "GSM1977746", "GSM1977747", "GSM1977748", "GSM1977749", "GSM1977750", "GSM1977751", "GSM1977752", "GSM1977753", "GSM1977754", "GSM1977755", "GSM1977756", "GSM1977757", "GSM1977758", "GSM1977759", "GSM1977760", "GSM1977761", "GSM1977762", "GSM1977763", "GSM1977764", "GSM1977765", "GSM1977766", "GSM1977767", "GSM1977768", "GSM1977769", "GSM1977770", "GSM1977771", "GSM1977772", "GSM1977773", "GSM1977774", "GSM1977775", "GSM1977776", "GSM1977777", "GSM1977778", "GSM1977779", "GSM1977780", "GSM1977781", "GSM1977782", "GSM1977783", "GSM1977784", "GSM1977785", "GSM1977786", "GSM1977787", "GSM1977788", "GSM1977789", "GSM1977790", "GSM1977791", "GSM1977792", "GSM1977793", "GSM1977794", "GSM1977795", "GSM1977796", "GSM1977797", "GSM1977798", "GSM1977799", "GSM1977800", "GSM1977801", "GSM1977802", "GSM1977803", "GSM1977804", "GSM1977805", "GSM1977806", "GSM1977807", "GSM1977808", "GSM1977809", "GSM1977810", "GSM1977811", "GSM1977812", "GSM1977813", "GSM1977814", "GSM1977815", "GSM1977816", "GSM1977817", "GSM1977818", "GSM1977819", "GSM1977820", "GSM1977821", "GSM1977822", "GSM1977823", "GSM1977824", "GSM1977825", "GSM1977826", "GSM1977827", "GSM1977828", "GSM1977829", "GSM1977830", "GSM1977831", "GSM1977832", "GSM1977833", "GSM1977834", "GSM1977835", "GSM1977836", "GSM1977837", "GSM1977838", "GSM1977839", "GSM1977840", "GSM1977841", "GSM1977842", "GSM1977843"], "sample_count": 198, "group_sizes": null, "group_sizes_basis": null, "platform": "17586", "taxon": "Homo sapiens", "title": "Comprehensive transcriptome profiling reveals multigene signatures in triple negative breast cancer", "resolution_checked_at": "2026-11-01T15:39:55Z", "resolution_check": "identifier_resolution_only_no_download_no_licence_read", "downloaded": false, "licence_read": false}}], "gaps": ["asset:data:geo:GSE58812 not downloaded", "asset:data:geo:GSE58812 licence not read", "asset:data:geo:GSE76250 not downloaded", "asset:data:geo:GSE76250 licence not read"]}, "analysis": {"state": "partial", "methods": [{"id": "method-1", "family": "machine_learning", "specific": "support_vector_machine", "label": "Machine learning", "dsl_operations": ["train_classifier", "cross_validate", "roc_auc"], "missing_operations": [], "expressible": true, "parameters": {}, "data_shapes": ["labelled_feature_matrix"], "evidence": {"score": 1.0, "band": "verbatim", "quote": "In the TCGA and GSE76250, to avoid possible influence of multicollinearity, we firstly conducted the LASSO logistic regression analysis through the “glmnet” package to screen out biomarkers from above all 60 CMGs at the optimal value of log lambda with the smallest classification error (30). Besides, the SVM-RFE machine learning algorithm based on the support vector machine was utilized to identify the most valuable biomarkers from all 60 CMGs by subtracting the feature vector determined using SVM with the “e1071” and “caret” R packages (31)."}}, {"id": "method-2", "family": "group_comparison", "specific": "mann_whitney", "label": "Group comparison", "dsl_operations": [], "missing_operations": ["group_comparison_mann_whitney"], "expressible": false, "parameters": {}, "data_shapes": ["two_or_more_groups_continuous"], "evidence": {"score": 0.8, "band": "high", "quote": "Mann-Whitney"}}, {"id": "method-3", "family": "principal_component_analysis", "specific": "pca", "label": "Principal component analysis", "dsl_operations": ["pca_fit"], "missing_operations": [], "expressible": true, "parameters": {}, "data_shapes": ["wide_numeric_matrix"], "evidence": {"score": 0.8, "band": "high", "quote": "principal component analysis"}}, {"id": "method-4", "family": "regression_modelling", "specific": "logistic", "label": "Regression modelling", "dsl_operations": ["glm_logit_fit"], "missing_operations": [], "expressible": true, "parameters": {}, "data_shapes": ["predictors_and_outcome"], "evidence": {"score": 0.8, "band": "high", "quote": "logistic regression"}}, {"id": "method-5", "family": "clustering", "specific": "kmeans", "label": "Clustering", "dsl_operations": ["kmeans_fit"], "missing_operations": [], "expressible": true, "parameters": {}, "data_shapes": ["feature_matrix"], "evidence": {"score": 0.7, "band": "medium", "quote": "K-means"}}], "targets": [], "gaps": ["method.reader_differs_from_extractor"]}, "capability": {"state": "complete", "requirements": [{"capability_id": "capability:computational_analysis", "label": "Computational analysis platform", "assay_families": ["meta_analysis", "machine_learning", "survival_analysis", "statistical_regression", "pca", "molecular_docking", "network_analysis", "differential_expression", "gwas"], "availability_state": "owned", "tier_unlocked": "R1", "candidate": "Automated Assays Laravel application and simulated adapters"}], "gaps": []}, "readiness": {"state": "complete", "tier": "R1", "blocking_axis": "none", "blocked_on": "nothing:replicable_now", "axes": {"method": {"tier": "R1", "expressible": true, "required_operations": ["train_classifier", "cross_validate", "roc_auc"], "missing_operations": [], "basis": "The paper names the method it used and the corpus's method extractor maps it onto the DSL catalogue: support_vector_machine."}, "capability": {"tier": "R1", "capability_id": "capability:computational_analysis", "availability_state": "owned", "basis": "The computational analysis is owned. The physical bench the paper used is not, so a physical validation cannot be run here."}, "data": {"tier": "R1", "reachable": true, "accessions": ["GSE76250", "GSE58812"], "resolved_accessions": ["GSE76250", "GSE58812"], "statements": ["The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material ."], "basis": "The paper names GSE76250, GSE58812 in Gene Expression Omnibus (GEO) and the accession resolves in the repository. No dataset was downloaded and no licence was read."}}, "can_run": {"dsl_operations": ["train_classifier", "cross_validate", "roc_auc"], "assets": ["source:pmid:39007147", "data:geo:GSE58812"], "steps": ["run_analysis"]}, "must_build": {"dsl_operations": [], "converters": [], "readers": [], "capabilities": [], "data_gaps": ["data:geo:GSE58812"]}, "gaps": []}, "review": {"state": "partial", "machine_critique": {"kind": "extraction_review", "reader": {"model": "MiniMax-M3.1-Flash", "provider": "minimax", "provider_family": "minimax", "read_at": "2026-11-01T19:24:02Z", "independence": "single_reader"}, "evidence_check": {"method": "verbatim", "availability": "verbatim"}, "calls": 5, "usage": {"prompt_tokens": 16094, "completion_tokens": 2118, "cached_tokens": 0}}, "human_critique": null, "stages": [{"id": "extraction", "state": "complete", "reader": "MiniMax-M3.1-Flash", "provider": "minimax"}], "gaps": ["review.independent_reader", "review.machine_critique", "review.human_critique"]}, "execution": {"state": "absent", "status": "not_run", "mode": null, "runs": [], "gaps": ["execution.not_run"]}, "provenance": {"entities": [{"id": "pmid:39007147", "type": "paper", "label": "pmid:39007147"}, {"id": "source:pmid:39007147", "type": "manuscript", "label": "PMID 39007147 full text"}, {"id": "data:geo:GSE58812", "type": "dataset", "label": "GEO GSE58812"}], "activities": [{"id": "activity-reading-sci-2026-11-01", "type": "classification", "used": [{"id": "pmid:39007147", "type": "paper", "label": "pmid:39007147"}, {"id": "source:pmid:39007147", "type": "manuscript", "label": "PMID 39007147 full text"}, {"id": "data:geo:GSE58812", "type": "dataset", "label": "GEO GSE58812"}], "agent": "agent:corpus-assay-engine"}, {"id": "activity-read-39007147", "type": "extraction", "used": ["source:pmid:39007147"], "agent": "agent:minimax-reader"}, {"id": "activity-resolve-GSE58812", "type": "accession_resolution", "used": ["data:geo:GSE58812"], "agent": "agent:corpus-assay-engine"}], "agents": [{"id": "agent:corpus-assay-engine", "type": "software", "label": "Corpus assay engine"}, {"id": "agent:minimax-reader", "type": "software", "label": "MiniMax literature reader"}], "source_run_id": "reading-sci-2026-11-01", "dsl_catalogue_fingerprint": "343fb30756ba6ec3", "capability_table_fingerprint": "2cb3e42f047d2ecf", "snapshot_id": "ba34b49490d9bfaf"}, "lineage": {"parent_id": "pmid:39007147", "parent_hash": "sha256:22f35de4b8d664bca6e21ee1df97358e418354fbdd24ff0907793f1d315e0a41", "fork_of": null, "generation": 3}, "gaps": ["asset:data:geo:GSE58812 licence not read", "asset:data:geo:GSE58812 not downloaded", "asset:data:geo:GSE76250 licence not read", "asset:data:geo:GSE76250 not downloaded", "execution.not_run", "method.reader_differs_from_extractor", "review.human_critique", "review.independent_reader", "review.machine_critique"], "assays": [{"role": "data_generation", "definition_id": "ihc", "definition_version": "99e1195c7d6f", "definition_digest": null, "required_operations": [], "support_state": "partial", "source_hashes": {"assay_lexicon": "99e1195c7d6f"}, "confidence": 0.99, "evidence": ["ing machine learning algorithms and multiple immunohistochemistry validation to unveil novel diagnostic marker", "response to immunotherapy via the multiplex immunohistochemistry (mIHC). Results: A total of 60 costimulatory"], "order_basis": "paper_text_order:8", "id": "assay:1"}, {"role": "analysis", "definition_id": "machine_learning", "definition_version": "f2132e96ae6c7a4e", "definition_digest": null, "required_operations": ["train_classifier", "cross_validate", "roc_auc"], "support_state": "supported", "source_hashes": {"dsl_catalogue": "f2132e96ae6c7a4e"}, "confidence": 1.0, "evidence": ["In the TCGA and GSE76250, to avoid possible influence of multicollinearity, we firstly conducted the LASSO logistic regression analysis through the “glmnet” package to screen out biomarkers from above all 60 CMGs at the optimal value of log lambda with the smallest classification error (30). Besides, the SVM-RFE machine learning algorithm based on the support vector machine was utilized to identify the most valuable biomarkers from all 60 CMGs by subtracting the feature vector determined using SVM with the “e1071” and “caret” R packages (31)."], "order_basis": "consumes_earlier_output", "id": "assay:2"}]}