Pan-Kanser Makine Öğrenimi Analizi Kullanılarak Tümör Sınıflandırması İçin Çekirdek Bir SASP Gen İmzasının Belirlenmesi

Öz

Bu çalışmada, literatürden türetilen ve yaygın olarak kullanılan 21 genlik bir SASP imzasının, makine öğrenimi (MÖ) modelleri kullanılarak birden fazla kanser türünde tümör ve normal dokuları ayırt edip edemeyeceğini test ettik. Meme kanseri (BRCA), akciğer adenokarsinomu (LUAD) ve kolon adenokarsinomu (COAD) RNA-seq verileri Kanser Genom Atlası'ndan (TCGA) elde edilmiştir. Sınıflandırma analizleri, Rastgele Orman, Destek Vektör Makineleri (SVM) ve XGBoost modelleri kullanılarak gerçekleştirilmiştir. Model kararlarını yorumlamak ve SASP imzası içindeki tahminleri etkileyen genleri belirlemek için SHAP (SHapley Additive exPlanations) tabanlı Açıklanabilir Yapay Zeka (XAI) analizi uygulanmıştır. Her bir makine öğrenimi modeli, yüksek performansla (AUC > 0,99) malign ve sağlıklı örnekleri doğru şekilde ayırdı. SHAP analizleri, tüm makine öğrenimi modellerinde sınıflandırma performansına en güçlü katkıyı sağlayan altı geni (MMP1, MMP3, PLAU, IL6, HGF ve EREG) ortak genler olarak vurgulamıştır. Fonksiyonel zenginleştirme analizleri, bu genleri hücre dışı matris yeniden şekillenmesi, PI3K-Akt sinyallemesi ve inflamasyon ile ilişkilendirdi. Modeller, GEO veri setleri kullanılarak harici doğrulamayla test edildi ve sonuçların ana bulguları desteklediği gözlemlenmiştir. Bu bulgular, kısa bir SASP türetilmiş imzanın kanserin genel moleküler parmak izi olarak işlev görebileceğini göstermektedir. Ayrıca, çalışma, biyobelirteç odaklı araştırmalar için yararlı olabilecek yorumlanabilir ve tekrarlanabilir gen modüllerini ortaya çıkarmak için açıklanabilir makine öğrenimi çerçevelerinin faydasını vurgulamaktadır.

Anahtar Kelimeler

• Kanser sınıflandırması
• Açıklanabilir yapay zeka
• Makine öğrenimi
• Yaşlanma ile ilişkili salgısal fenotip
• Biyobelirteç önceliklendirme

Etik Beyan

Bu makalenin yazarı, bu çalışmada kullanılan materyaller ve yöntemlerin etik kurul izni ve/veya özel yasal izin gerektirmediğini beyan etmektedir.

Identification of a Core SASP Gene Signature for Tumor Classification Using Pan-Cancer Machine Learning Analysis

Öz

In this study, we tested whether a literature-derived 21-genes commonly used SASP signature to discriminate tumor and normal tissues across multiple cancer types using machine learning (ML) models. RNA-seq data from breast cancer (BRCA), lung adenocarcinoma (LUAD), and colon adenocarcinoma (COAD) were obtained from The Cancer Genome Atlas (TCGA). Classification analyses were performed using Random Forest, Support Vector Machines (SVM), and XGBoost models. Explainable Artificial Intelligence (XAI) analysis based on SHAP (SHapley Additive exPlanations) was applied to interpret model decisions and identify genes influencing predictions within the SASP signature. Each ML model correctly separated malignant from healthy samples with high performance (AUC > 0.99). SHAP analyses highlighted six genes (MMP1, MMP3, PLAU, IL6, HGF, and EREG) as mutual genes in all ML models that contributed at most strongly to classification performance. Functional enrichment analyses linked these genes to extracellular-matrix remodeling, PI3K-Akt signaling, and inflammation. The models were tested with external validation using GEO datasets, and the results supported the main findings. These findings suggest that a concise SASP-derived signature can act as a general molecular fingerprint of cancer. Moreover, the study highlights the utility of explainable machine learning frameworks for uncovering interpretable and reproducible gene modules which may be useful for biomarker-oriented research.

Anahtar Kelimeler

• Cancer classification
• Explainable artificial intelligence
• Machine learning
• Senescence-associated secretory phenotype
• Biomarker prioritization

Etik Beyan

The author of this article declares that the materials and methods used in this study do not require ethical committee permission and/or legal-special permission.

Kaynakça

1. [1] Hernandez-Segura A., Nehme J. and Demaria M., “Hallmarks of cellular senescence”, Trends in Cell Biology, 28(6): 436–453, (2018).
2. [2] Chaib S., Tchkonia T. and Kirkland J.L., “Cellular senescence and senolytics: the path to the clinic”, Nature Medicine, 28: 1556–1568, (2022).
3. [3] Gorgoulis V.G., Adams P.D., Alimonti A., Bennett D.C., Bischof O., Bishop C. et al., “Cellular senescence: Defining a path forward”, Cell, 179(4): 813–827, (2019).
4. [4] Ajoolabady A., Pratico D., Bahijri S. et al., “Hallmarks and mechanisms of cellular senescence in aging and disease”, Cell Death Discovery, 11: 364, (2025).
5. [5] Coppé J.P., Desprez P.Y., Krtolica A. and Campisi J., “The senescence-associated secretory phenotype: The dark side of tumor suppression”, Annual Review of Pathology, 5: 99–118, (2010).
6. [6] Basisty N., Kale A., Jeon O.H., Kuehnemann C., Payne T., Rao C. et al., “A proteomic atlas of senescence-associated secretomes for aging biomarker development”, PLOS Biology, 18(1): e3000599, (2020).
7. [7] Faget D.V., Ren Q. and Stewart S.A., “Unmasking senescence: Context-dependent effects of SASP in cancer”, Nature Reviews Cancer, 19(8): 439–453, (2019).
8. [8] Dong Z., Luo Y., Yuan Z., Tian Y., Jin T. and Xu F., “Cellular senescence and SASP in tumor progression and therapeutic opportunities”, Molecular Cancer, 23: 181, (2024).

9. [9] Duran I., Pombo J., Sun B., Gallage S., Kudo H., McHugh D. et al., “Detection of senescence using machine learning algorithms based on nuclear features”, Nature Communications, 15: 1041, (2024).
10. [10] Zhang X., Rahnavard A. and Crandall K.A., “Machine learning enhances biomarker discovery: From multi-omics to functional genomics”, Medical Research Archives, 13(8), (2025).
11. [11] Fountzilas E., Pearce T., Baysal M.A., Chakraborty A. and Tsimberidou A.M., “Convergence of evolving artificial intelligence and machine learning techniques in precision oncology”, npj Digital Medicine, 8: 75, (2025).
12. [12] Lundberg S.M. and Lee S.I., “A unified approach to interpreting model predictions”, Advances in Neural Information Processing Systems, 30, (2017).
13. [13] Hanani A.A., Donmez T.B., Kutlu M., Mansour M., “Predicting thyroid cancer recurrence using supervised CatBoost: A SHAP-based explainable AI approach”, Medicine (Baltimore), 104(22): e42667, (2025).
14. [14] Donmez T.B., Kutlu M., Mansour M. et al., “Explainable AI in action: A comparative analysis of hypertension risk factors using SHAP and LIME”, Neural Computing and Applications, 37: 4053–4074, (2025).
15. [15] SeneQuest Database, “A comprehensive database of senescence-associated genes”, Available at: https://senequest.net, (accessed 04 October 2025).
16. [16] TCGA Research Network, “The Cancer Genome Atlas Program”, Available at: https://www.cancer.gov/tcga, (accessed 04 October 2025).
17. [17] Colaprico A., Silva T.C., Olsen C., Garofano L., Cava C., Garolini D. et al., “TCGAbiolinks: An R/Bioconductor package for integrative analysis of TCGA data”, Nucleic Acids Research, 44(8): e71, (2016).
18. [18] Pedregosa F., Varoquaux G., Gramfort A., Michel V., Thirion B., Grisel O. et al., “Scikit-learn: Machine learning in Python”, Journal of Machine Learning Research, 12: 2825–2830, (2011).
19. [19] Chen T. and Guestrin C., “XGBoost: A scalable tree boosting system”, Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 785–794, (2016).
20. [20] Yu G., Wang L.G., Han Y. and He Q.Y., “clusterProfiler: An R package for comparing biological themes among gene clusters”, OMICS, 16(5): 284–287, (2012).
21. [21] Carlson M., “org.Hs.eg.db: Genome wide annotation for Human”, R package version 3.17.0, Bioconductor, (2019).
22. [22] Ashburner M., Ball C.A., Blake J.A., Botstein D., Butler H., Cherry J.M. et al., “Gene ontology: Tool for the unification of biology”, Nature Genetics, 25(1): 25–29, (2000).
23. [23] Kanehisa M. and Goto S., “KEGG: Kyoto encyclopedia of genes and genomes”, Nucleic Acids Research, 28(1): 27–30, (2000).
24. [24] Benjamini Y. and Hochberg Y., “Controlling the false discovery rate: A practical and powerful approach to multiple testing”, Journal of the Royal Statistical Society: Series B, 57(1): 289–300, (1995).
25. [25] Venet D., Dumont J.E. and Detours V., “Most random gene expression signatures are significantly associated with breast cancer outcome”, PLOS Computational Biology, 7(10): e1002240, (2011).
26. [26] Edgar R., Domrachev M. and Lash A.E., “Gene Expression Omnibus: NCBI gene expression and hybridization array data repository”, Nucleic Acids Research, 30(1): 207–210, (2002).
27. [27] Barrett T., Wilhite S.E., Ledoux P., Evangelista C., Kim I.F., Tomashevsky M. et al., “NCBI GEO: archive for functional genomics data sets—update”, Nucleic Acids Research, 41(Database issue): D991–D995, (2013).
28. [28] Seo J.S., Ju Y.S., Lee W.C., Shin J.Y., Lee J.K., Bleazard T. et al., “The transcriptional landscape and mutational profile of lung adenocarcinoma”, Genome Research, 22(11): 2109–2119, (2012).
29. [29] Wenric S., ElGuendi S., Caberg J.H., Bezzaou W., Fasquelle C., Charloteaux B. et al., “Transcriptome-wide analysis of natural antisense transcripts shows their potential role in breast cancer”, Scientific Reports, 7: 17452, (2017).
30. [30] Kim S.K., Kim S.Y., Kim J.H., Roh S.A., Cho D.H., Kim Y.S. et al., “A nineteen gene-based risk score classifier predicts prognosis of colorectal cancer patients”, Molecular Oncology, 8(8): 1653–1666, (2014).
31. [31] Kim J.C., Ha Y.J., Tak K.H., Roh S.A., Kwon Y.H., Kim Y.S. et al., “Complex behavior of ALDH1A1 and IGFBP1 in liver metastasis from a colorectal cancer”, PLOS ONE, 11(5): e0155160, (2016).
32. [32] Kim S.K., Kim S.Y., Kim C.W., Roh S.A., Cho D.H., Kim Y.S. et al., “A prognostic index based on an eleven gene signature to predict systemic recurrences in colorectal cancer”, Experimental & Molecular Medicine, 51(10): 1–12, (2019).
33. [33] Ha Y.J., Park S.H., Tak K.H., Lee J.L., Roh S.A., Kim Y.S. et al., “CILP2 is a potential biomarker for the prediction and therapeutic target of peritoneal metastases in colorectal cancer”, Scientific Reports, 14: 12487, (2024).
34. [34] Szklarczyk D., Kirsch R., Koutrouli M. et al., “The STRING database in 2023: protein–protein association networks and functional enrichment analyses for any sequenced genome of interest”, Nucleic Acids Research, 51(D1): D638–D646, (2023).
35. [35] Lonsdale J., Thomas J., Salvatore M., Phillips R., Lo E., Shad S. et al., “The Genotype-Tissue Expression (GTEx) project”, Nature Genetics, 45(6): 580–585, (2013).
36. [36] GTEx Consortium, “Human genomics. The Genotype-Tissue Expression (GTEx) pilot analysis: multitissue gene regulation in humans”, Science, 348(6235): 648–660, (2015).
37. [37] Egeblad M. and Werb Z., “New functions for the matrix metalloproteinases in cancer progression”, Nature Reviews Cancer, 2(3): 161–174, (2002).
38. [38] Andreasen P.A., Kjoller L., Christensen L. and Duffy M.J., “The urokinase-type plasminogen activator system in cancer metastasis: A review”, International Journal of Cancer, 72(1): 1–22, (1998).
39. [39] Duffy M.J., McGowan P.M. and Gallagher W.M., “Cancer invasion and metastasis: Changing views”, Journal of Pathology, 214(3): 283–293, (2008).
40. [40] Trusolino L., Bertotti A. and Comoglio P.M., “MET signalling: Principles and functions in development, organ regeneration and cancer”, Nature Reviews Molecular Cell Biology, 11(12): 834–848, (2010).
41. [41] Normanno N., De Luca A., Bianco C., Strizzi L., Mancino M., Maiello M.R. et al., “Epidermal growth factor receptor (EGFR) signaling in cancer”, Gene, 366(1): 2–16, (2006).
42. [42] [Hanahan D. and Weinberg R.A., “Hallmarks of cancer: The next generation”, Cell, 144(5): 646–674, (2011).
43. [43] Uhlén M., Fagerberg L., Hallström B.M. et al., “Proteomics: Tissue-based map of the human proteome”, Science, 347(6220): 1260419, (2015).
44. [44] Uhlén M., Karlsson M.J., Zhong W. et al., “A genome-wide transcriptomic analysis of protein-coding genes in human blood cells”, Science, 366(6472): eaax9198, (2019).