Ablation studies and performance evaluation under varying training conditions.
<p>(A) Performance comparison under four topological feature integration settings: no features (None), only global features (w/o local), only local features (w/o global), and both local and global features (Full). The full model achieves the best results across ACC, AUROC, and AUPRC metrics, c...
Պահպանված է:
| Հիմնական հեղինակ: | |
|---|---|
| Այլ հեղինակներ: | , , |
| Հրապարակվել է: |
2025
|
| Խորագրեր: | |
| Ցուցիչներ: |
Ավելացրեք ցուցիչ
Չկան պիտակներ, Եղեք առաջինը, ով նշում է այս գրառումը!
|
| _version_ | 1849927641103073280 |
|---|---|
| author | Fatemeh Keikha (20917848) |
| author2 | Chuanyuan Wang (22676721) Zhixia Yang (3180567) Zhi-Ping Liu (149304) |
| author2_role | author author author |
| author_facet | Fatemeh Keikha (20917848) Chuanyuan Wang (22676721) Zhixia Yang (3180567) Zhi-Ping Liu (149304) |
| author_role | author |
| dc.creator.none.fl_str_mv | Fatemeh Keikha (20917848) Chuanyuan Wang (22676721) Zhixia Yang (3180567) Zhi-Ping Liu (149304) |
| dc.date.none.fl_str_mv | 2025-11-24T18:36:51Z |
| dc.identifier.none.fl_str_mv | 10.1371/journal.pcbi.1013743.g004 |
| dc.relation.none.fl_str_mv | https://figshare.com/articles/figure/Ablation_studies_and_performance_evaluation_under_varying_training_conditions_/30697984 |
| dc.rights.none.fl_str_mv | CC BY 4.0 info:eu-repo/semantics/openAccess |
| dc.subject.none.fl_str_mv | Medicine Genetics Biotechnology Cancer Infectious Diseases Biological Sciences not elsewhere classified Mathematical Sciences not elsewhere classified Information Systems not elsewhere classified wasserstein optimal transport real world dataset quantified via gromov enhancing diagnostic precision dynamic network index deep neural network ablation studies confirm regulatory role transitions gene regulatory networks combining regulatory rewiring analyzing disease evolution temporal expression dynamics reflect dynamic changes neglecting structural rewiring multilayer network models modeling disease progression state alignment provides understanding cancer progression state graph alignment disease state classification state specific biomarkers disease progression state alignment specific expression multilayer graph cancer progression disease state regulatory variability regulatory roles structural shifts significant changes prioritized biomarkers classification accuracy state single xlink "> topological features synthetic simulated ranked using powerful strategy overall performance meaningful shifts gats ), gastric adenocarcinoma gac ), framework designed findings suggest dni ), distinct layer contextualized embeddings cell data |
| dc.title.none.fl_str_mv | Ablation studies and performance evaluation under varying training conditions. |
| dc.type.none.fl_str_mv | Image Figure info:eu-repo/semantics/publishedVersion image |
| description | <p>(A) Performance comparison under four topological feature integration settings: no features (None), only global features (w/o local), only local features (w/o global), and both local and global features (Full). The full model achieves the best results across ACC, AUROC, and AUPRC metrics, confirming the complementary value of integrating both feature types. (B) Comparison of different network rewiring strategies: no rewiring, correlation-based rewiring, our proposed regulatory rewiring, and rewiring with random noise injection (10%, 20%, and 30%). Our method outperforms others, demonstrating both the importance of state-specific regulatory rewiring and the robustness of the approach under noisy conditions. (C–D) Sensitivity analysis of the balancing coefficient <i>α</i> that controls the contribution of local versus global topological features. Optimal performance is achieved at <i>α</i>=0.3, highlighting the importance of a balanced integration. (E–F) Evaluation of model performance with varying training data sizes (10% to 50%). AUROC box plots (E) and full performance metrics (F) show that larger training sets yield more stable and accurate predictions. (G) Evaluation of model performance with different embedding dimensionalities (32 to 512) based on ACC, AUROC, and AUPRC. (H) Performance of the model under different dropout rates. (I) Classification performance using varying proportions (1% to 30%) of top-ranked genes; optimal performance is achieved using the top 15%.</p> |
| eu_rights_str_mv | openAccess |
| id | Manara_a359ed4a7cf37c8227371baa31cc77fb |
| identifier_str_mv | 10.1371/journal.pcbi.1013743.g004 |
| network_acronym_str | Manara |
| network_name_str | ManaraRepo |
| oai_identifier_str | oai:figshare.com:article/30697984 |
| publishDate | 2025 |
| repository.mail.fl_str_mv | |
| repository.name.fl_str_mv | |
| repository_id_str | |
| rights_invalid_str_mv | CC BY 4.0 |
| spelling | Ablation studies and performance evaluation under varying training conditions.Fatemeh Keikha (20917848)Chuanyuan Wang (22676721)Zhixia Yang (3180567)Zhi-Ping Liu (149304)MedicineGeneticsBiotechnologyCancerInfectious DiseasesBiological Sciences not elsewhere classifiedMathematical Sciences not elsewhere classifiedInformation Systems not elsewhere classifiedwasserstein optimal transportreal world datasetquantified via gromovenhancing diagnostic precisiondynamic network indexdeep neural networkablation studies confirmregulatory role transitionsgene regulatory networkscombining regulatory rewiringanalyzing disease evolutiontemporal expression dynamicsreflect dynamic changesneglecting structural rewiringmultilayer network modelsmodeling disease progressionstate alignment providesunderstanding cancer progressionstate graph alignmentdisease state classificationstate specific biomarkersdisease progressionstate alignmentspecific expressionmultilayer graphcancer progressiondisease stateregulatory variabilityregulatory rolesstructural shiftssignificant changesprioritized biomarkersclassification accuracystate singlexlink ">topological featuressynthetic simulatedranked usingpowerful strategyoverall performancemeaningful shiftsgats ),gastric adenocarcinomagac ),framework designedfindings suggestdni ),distinct layercontextualized embeddingscell data<p>(A) Performance comparison under four topological feature integration settings: no features (None), only global features (w/o local), only local features (w/o global), and both local and global features (Full). The full model achieves the best results across ACC, AUROC, and AUPRC metrics, confirming the complementary value of integrating both feature types. (B) Comparison of different network rewiring strategies: no rewiring, correlation-based rewiring, our proposed regulatory rewiring, and rewiring with random noise injection (10%, 20%, and 30%). Our method outperforms others, demonstrating both the importance of state-specific regulatory rewiring and the robustness of the approach under noisy conditions. (C–D) Sensitivity analysis of the balancing coefficient <i>α</i> that controls the contribution of local versus global topological features. Optimal performance is achieved at <i>α</i>=0.3, highlighting the importance of a balanced integration. (E–F) Evaluation of model performance with varying training data sizes (10% to 50%). AUROC box plots (E) and full performance metrics (F) show that larger training sets yield more stable and accurate predictions. (G) Evaluation of model performance with different embedding dimensionalities (32 to 512) based on ACC, AUROC, and AUPRC. (H) Performance of the model under different dropout rates. (I) Classification performance using varying proportions (1% to 30%) of top-ranked genes; optimal performance is achieved using the top 15%.</p>2025-11-24T18:36:51ZImageFigureinfo:eu-repo/semantics/publishedVersionimage10.1371/journal.pcbi.1013743.g004https://figshare.com/articles/figure/Ablation_studies_and_performance_evaluation_under_varying_training_conditions_/30697984CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/306979842025-11-24T18:36:51Z |
| spellingShingle | Ablation studies and performance evaluation under varying training conditions. Fatemeh Keikha (20917848) Medicine Genetics Biotechnology Cancer Infectious Diseases Biological Sciences not elsewhere classified Mathematical Sciences not elsewhere classified Information Systems not elsewhere classified wasserstein optimal transport real world dataset quantified via gromov enhancing diagnostic precision dynamic network index deep neural network ablation studies confirm regulatory role transitions gene regulatory networks combining regulatory rewiring analyzing disease evolution temporal expression dynamics reflect dynamic changes neglecting structural rewiring multilayer network models modeling disease progression state alignment provides understanding cancer progression state graph alignment disease state classification state specific biomarkers disease progression state alignment specific expression multilayer graph cancer progression disease state regulatory variability regulatory roles structural shifts significant changes prioritized biomarkers classification accuracy state single xlink "> topological features synthetic simulated ranked using powerful strategy overall performance meaningful shifts gats ), gastric adenocarcinoma gac ), framework designed findings suggest dni ), distinct layer contextualized embeddings cell data |
| status_str | publishedVersion |
| title | Ablation studies and performance evaluation under varying training conditions. |
| title_full | Ablation studies and performance evaluation under varying training conditions. |
| title_fullStr | Ablation studies and performance evaluation under varying training conditions. |
| title_full_unstemmed | Ablation studies and performance evaluation under varying training conditions. |
| title_short | Ablation studies and performance evaluation under varying training conditions. |
| title_sort | Ablation studies and performance evaluation under varying training conditions. |
| topic | Medicine Genetics Biotechnology Cancer Infectious Diseases Biological Sciences not elsewhere classified Mathematical Sciences not elsewhere classified Information Systems not elsewhere classified wasserstein optimal transport real world dataset quantified via gromov enhancing diagnostic precision dynamic network index deep neural network ablation studies confirm regulatory role transitions gene regulatory networks combining regulatory rewiring analyzing disease evolution temporal expression dynamics reflect dynamic changes neglecting structural rewiring multilayer network models modeling disease progression state alignment provides understanding cancer progression state graph alignment disease state classification state specific biomarkers disease progression state alignment specific expression multilayer graph cancer progression disease state regulatory variability regulatory roles structural shifts significant changes prioritized biomarkers classification accuracy state single xlink "> topological features synthetic simulated ranked using powerful strategy overall performance meaningful shifts gats ), gastric adenocarcinoma gac ), framework designed findings suggest dni ), distinct layer contextualized embeddings cell data |