Review

Trends in computational approaches for anti-cancer research

Hanadi M. Baeissa ^1*

¹ Department of Biochemistry, College of Science, University of Jeddah, Jeddah-21589 Saudi Arabia.

* Correspondence: hmbaeissa@uj.edu.sa (H.M.B.)

---

Citation: Baeissa H.M. Trends in computational approaches for anti-cancer research. Glob. Jour. Bas. Sci. 2025, 1(8). 1-8.

Received: March 11, 2025

Revised: June 19, 2025

Accepted: June 22, 2025

Published: June 24, 2025

doi: 10.63454/jbs20000039

ISSN: 3049-3315

Volume 1; Issue 8

Download PDF file

---

Abstract: Computational approaches have become indispensable in anti-cancer research, enabling efficient analysis of complex biological data, prediction of therapeutic targets, and rational drug design. This commentary reviews current trends in computational oncology, including machine learning, network biology, structural bioinformatics, systems pharmacology, and multi-omics integration. It highlights major applications such as biomarker discovery, drug repurposing, precision oncology, and simulation of tumor evolution. We discuss challenges, including data heterogeneity, interpretability, and reproducibility, as well as emerging directions in artificial intelligence, cloud computing, and digital pathology. Integrating computational strategies with experimental validation is essential for accelerating translational impact and improving clinical outcomes.

Keywords: computational oncology; machine learning; network biology; systems pharmacology; multi-omics; bioinformatics; drug discovery; precision medicine

1. Introduction

Cancer remains one of the leading causes of mortality worldwide, characterized by genetic heterogeneity, dynamic evolution, and complex tumor microenvironments. Traditional laboratory methods have generated vast datasets but often lack the scalability to uncover high-order biological relationships. Computational approaches, leveraging advances in computer science, bioinformatics, and systems biology, have emerged as powerful tools to address these challenges by enabling integration, analysis, and interpretation of high-dimensional biological data [1,2]. These methods facilitate a deeper understanding of cancer biology, support novel target identification, accelerate drug discovery, and drive precision oncology.  This commentary synthesizes the latest trends in computational approaches applied to anti-cancer research, surveys key methodologies, and discusses their impact on translational science and clinical practice.

2. Machine learning and artificial intelligence in cancer research

2.1 Predictive modeling and diagnosis

Machine learning (ML) and artificial intelligence (AI) have revolutionized cancer research by enabling the development of predictive models that can diagnose disease, stratify patients, and forecast clinical outcomes with remarkable accuracy (Figure 1). Traditional supervised learning algorithms such as random forests, support vector machines (SVMs), and k-nearest neighbors (k-NN) have been widely applied to classify tumor subtypes based on gene expression profiles, proteomic signatures, and epigenetic markers. These models have  demonstrated high sensitivity and specificity in distinguishing between malignant and benign lesions, as well as identifying clinically relevant subgroups of cancers [3,4].

Recent advances in deep learning, particularly convolutional neural networks (CNNs) and recurrent neural networks (RNNs), have further expanded diagnostic capabilities. CNNs excel in image-based tasks, enabling automated detection of tumors in histopathological slides, radiographic scans, and endoscopic images. RNNs and transformer-based architectures have been applied to genomic and transcriptomic data, uncovering subtle sequence patterns that may escape human expertise. These models not only improve diagnostic precision but also support early detection, which is critical for improving survival rates in cancers such as lung, breast, and colorectal malignancies [5].

Figure 1. Machine learning in cancer research.

2.2 Feature selection and biomarker discovery

One of the most impactful contributions of ML in oncology is its ability to enhance biomarker discovery. Cancer datasets are inherently high-dimensional, with thousands of genes, proteins, and epigenetic modifications measured simultaneously. ML techniques such as LASSO regression, ridge regression, and elastic net are particularly effective in reducing dimensionality while retaining predictive features. These methods identify gene signatures and molecular markers associated with prognosis, therapeutic response, or resistance mechanisms [6].

Ensemble learning approaches, including random forests and gradient boosting machines, further improve biomarker discovery by aggregating predictions from multiple models, thereby increasing robustness and reducing variance. Multi-omics integration—combining genomics, transcriptomics, proteomics, and metabolomics—has been facilitated by ML frameworks that can uncover cross-layered interactions. This integrative approach enables the identification of composite biomarkers that provide a more holistic view of tumor biology and patient-specific vulnerabilities [7]. Such biomarkers are increasingly being incorporated into clinical trials and precision medicine initiatives, guiding targeted therapies and personalized treatment regimens.

2.3 Challenges in AI Implementation

Despite these advances, several challenges hinder the widespread clinical adoption of AI in oncology. Overfitting remains a major concern, particularly when models are trained on small or unbalanced datasets. Overfitted models may perform well on training data but fail to generalize to independent cohorts, limiting their clinical utility.

Another critical issue is interpretability. Many deep learning models function as “black boxes,” providing predictions without clear explanations of underlying mechanisms. This lack of transparency reduces clinician trust and complicates regulatory approval. Efforts to develop explainable AI (XAI) are underway, aiming to provide interpretable outputs such as feature importance scores, saliency maps, and decision pathways.

Bias in training data also poses ethical and practical challenges. If datasets underrepresent certain populations—such as ethnic minorities or patients from low-resource settings—models may yield inequitable predictions, exacerbating disparities in cancer care. Addressing this requires diverse, representative datasets and rigorous fairness assessments.

Finally, validation across independent cohorts is essential to ensure generalizability. Multi-center studies, external benchmarking, and prospective clinical trials are needed to confirm the robustness of AI models before they can be integrated into routine practice [8]. Overcoming these challenges will be critical to translating computational advances into tangible clinical benefits. 

3. Network biology and systems approaches

3.1 Biological networks in cancer

Network biology conceptualizes biological systems as interconnected units rather than isolated components, emphasizing the dynamic interplay among genes, proteins, and signaling pathways. In cancer, protein–protein interaction (PPI) networks reveal how oncogenic proteins form complexes that regulate cell cycle progression, apoptosis, and DNA repair (Figure 2). Similarly, signaling pathways such as PI3K/AKT/mTOR, MAPK, and JAK/STAT are modeled as interconnected cascades, highlighting cross-talk and redundancy that sustain tumorigenesis. Gene regulatory networks (GRNs), constructed from transcriptomic data, uncover transcription factors and non-coding RNAs that orchestrate malignant phenotypes. By mapping these networks, researchers can identify functional dependencies and emergent properties that drive cancer initiation, progression, and metastasis [9,10].

Figure 2. Network biology in cancer.

3.2 Network-based target prioritization

Network analysis enables the identification of central (“hub”) nodes—genes or proteins with high connectivity—that are critical for tumor survival. Targeting these hubs can disrupt multiple downstream processes, offering more effective therapeutic strategies than single-gene approaches [11]. Beyond hub identification, network medicine integrates disease networks, drug-target networks, and gene interaction maps to uncover opportunities for drug repurposing. For example, repositioning cardiovascular drugs with network-predicted anticancer activity has shown promise in preclinical studies [12]. This approach accelerates therapeutic discovery while reducing costs and timelines.

3.3 Modeling tumor evolution

Computational frameworks simulate tumor evolution and clonal dynamics, providing insights into how cancer adapts under selective pressures such as therapy. Models incorporating mutation rates, fitness landscapes, and microenvironmental interactions can predict the emergence of resistant clones and relapse mechanisms [13]. These simulations inform adaptive treatment strategies, such as alternating drug regimens or combination therapies, designed to preempt resistance and prolong patient survival.

4. Structural bioinformatics and rational drug design

4.1 Molecular docking and virtual screening

Molecular docking predicts the binding orientation and affinity of small molecules to target proteins, serving as a cornerstone of in silico drug discovery. By screening chemical libraries against oncogenic proteins, researchers can rapidly identify candidate compounds with favorable binding profiles. Virtual high-throughput screening (vHTS) reduces experimental burden by prioritizing compounds for laboratory validation, thereby accelerating the drug discovery pipeline [14,15].

4.2 Molecular dynamics and binding free energy

While docking provides static snapshots, molecular dynamics (MD) simulations refine predictions by modeling protein flexibility, solvent effects, and conformational changes over time. MD simulations allow calculation of binding free energies, improving accuracy in ranking ligands and predicting stability of drug–target complexes [16]. This dynamic perspective is crucial for  understanding how mutations alter drug binding and contribute to resistance.

4.3 Drug repurposing

Computational screening has facilitated the repurposing of FDA-approved drugs for oncology, leveraging existing safety profiles to reduce development time and cost. Examples include the identification of antidiabetic and anti-inflammatory drugs with anticancer activity. Network-based and structure-guided repurposing approaches expand the therapeutic arsenal by uncovering unexpected drug–target relationships [17].

5. Multi-omics integration

5.1 Integrative genomic analysis

Cancer is a multi-layered disease, and multi-omics integration—combining genomics, transcriptomics, proteomics, and epigenomics—provides a systems-level understanding of its biology. Integrative analyses reveal cross-layered interactions, such as how mutations in chromatin remodelers alter transcriptional programs and downstream proteomic signatures. These approaches uncover mechanistic drivers of malignancy, enabling precise stratification of patients and identification of novel therapeutic targets (18,19).

5.2 Challenges in data integration

Despite its promise, multi-omics integration faces challenges including data heterogeneity, batch effects, and normalization issues. Differences in platforms, sample preparation, and sequencing depth complicate cross-study comparisons. Advanced computational frameworks such as matrix factorization, Bayesian hierarchical models, and graph-based integration help harmonize datasets, improving robustness and reproducibility [20,21]. Continued methodological innovation is essential to fully exploit the potential of multi-omics data.

6. Systems pharmacology

6.1 Pathway-Based Drug Efficacy Prediction

Systems pharmacology integrates network biology with pharmacokinetics/pharmacodynamics (PK/PD) models to predict drug efficacy and toxicity. By simulating drug effects across signaling pathways, these models enable rational design of combination therapies that maximize efficacy while minimizing adverse effects [22]. Pathway-based predictions also help identify biomarkers for patient stratification, supporting precision medicine (Figure 3).

Figure 3. Systems pharmacological approach in anticancer research.

6.2 Synthetic lethality and combination therapy

Computational approaches can identify synthetic lethal interactions, where simultaneous disruption of two genes leads to cell death, while inhibition of either gene alone is tolerated. Exploiting synthetic lethality allows selective targeting of cancer cells with specific mutations. For example, PARP inhibitors exploit synthetic lethality in BRCA-mutated tumors. Network-guided identification of such interactions informs rational combination therapy design, offering new avenues to overcome resistance [23].

7. Computational pathology and digital imaging

7.1 AI in histopathology

Advances in deep learning have revolutionized histopathology by enabling automated interpretation of tissue images. Convolutional neural networks (CNNs) can detect tumors, grade malignancies, and quantify biomarkers with accuracy comparable to expert pathologists. These tools reduce diagnostic variability and enhance throughput, supporting clinical decision-making [24].

7.2 Integration with genomics

The emerging field of radiogenomics links imaging features with genomic and transcriptomic profiles, offering non-invasive biomarkers for cancer diagnosis and prognosis. For instance, imaging-derived features such as texture or vascularity can correlate with mutational status or gene expression signatures. Integrating digital pathology with genomics enhances predictive modeling, enabling personalized treatment strategies without invasive biopsies [25].

8. Cloud computing, big data, and reproducibility

The advent of large-scale cancer genomics initiatives such as The Cancer Genome Atlas (TCGA) and the International Cancer Genome Consortium (ICGC) has generated unprecedented volumes of multi-omics data, encompassing genomics, transcriptomics, epigenomics, proteomics, and clinical metadata. Analyzing these datasets requires robust computational infrastructures capable of handling high-throughput workflows, complex statistical models, and integrative pipelines. Traditional local computing environments often lack the scalability and storage capacity needed for such tasks, leading to the widespread adoption of cloud computing platforms (e.g., Amazon Web Services, Google Cloud, Microsoft Azure). These platforms provide elastic resources, enabling researchers to scale analyses dynamically and collaborate across institutions without the constraints of physical hardware.

Equally important is the issue of reproducibility, a cornerstone of scientific integrity. Containerization technologies such as Docker and Singularity allow researchers to encapsulate software environments, dependencies, and workflows into portable units. This ensures that analyses can be replicated across different systems, reducing variability introduced by operating systems or library versions. Workflow management systems (e.g., Nextflow, Snakemake) further enhance reproducibility by automating complex pipelines and tracking provenance. Together, cloud computing and containerization have transformed computational oncology into a more collaborative, transparent, and reproducible discipline.

9. Ethical, legal, and social considerations

The integration of computational methods into oncology raises profound ethical, legal, and social challenges. One of the foremost concerns is data privacy, as genomic and clinical datasets contain sensitive personal information that could be misused if not adequately protected. Ensuring compliance with regulations such as the General Data Protection Regulation (GDPR) and Health Insurance Portability and Accountability Act (HIPAA) is essential for safeguarding patient confidentiality.

Another critical issue is algorithmic bias. Machine learning models trained on datasets that underrepresent certain populations may yield inequitable predictions, exacerbating disparities in cancer diagnosis and treatment. Addressing bias requires diverse, representative datasets and transparent reporting of model limitations. Furthermore, equitable access to computational technologies remains a challenge, particularly in low- and middle-income countries where infrastructure and expertise may be limited. Without deliberate efforts, computational oncology risks widening the gap between resource-rich and resource-poor settings.

To mitigate these concerns, the establishment of ethical frameworks and transparent reporting standards is imperative. Initiatives such as FAIR (Findable, Accessible, Interoperable, Reusable) data principles and open-source software development promote accountability and inclusivity. Ultimately, responsible deployment of computational oncology tools must balance innovation with patient rights, equity, and societal trust.

10. Future trends and emerging technologies

10.1 Explainable AI

While artificial intelligence has demonstrated remarkable predictive power in oncology, its adoption in clinical practice is hindered by the “black box” nature of many models. Explainable AI (XAI) seeks to bridge this gap by developing interpretable algorithms that provide insights into how predictions are made. Techniques such as attention mechanisms, feature importance scoring, and visualization of decision pathways can enhance clinician trust and facilitate regulatory approval. By making AI outputs transparent, XAI will accelerate integration into diagnostic, prognostic, and therapeutic workflows.

10.2 Real-time analytics and wearables

The proliferation of wearable devices and mobile health technologies offers new opportunities for continuous monitoring of patient physiology, activity, and biometrics. Integrating real-time data streams with computational oncology models could enable early detection of relapse, dynamic monitoring of treatment responses, and personalized lifestyle interventions. For example, wearable-derived data on heart rate variability or sleep patterns may complement molecular biomarkers, providing a holistic view of patient health. Real-time analytics pipelines will be critical to process these data efficiently and translate them into actionable clinical insights.

10.3 Single-cell and spatial transcriptomics

Recent advances in single-cell sequencing and spatial transcriptomics are revolutionizing cancer biology by uncovering intratumor heterogeneity and microenvironmental interactions at unprecedented resolution. Computational methods tailored to single-cell data allow researchers to reconstruct lineage trajectories, identify rare subpopulations, and map cell–cell communication networks. Spatial transcriptomics adds a crucial dimension by preserving tissue architecture, enabling the study of how tumor cells interact with stromal and immune components. These technologies promise to refine our understanding of tumor evolution and resistance mechanisms, ultimately guiding the development of more precise therapeutic strategies.

11. Conclusion

Computational approaches have profoundly reshaped the landscape of anti-cancer research, enabling the integration and interpretation of vast, multidimensional biological datasets that were previously intractable. By harnessing advances in bioinformatics, machine learning, and systems biology, researchers can now uncover hidden molecular patterns, predict therapeutic responses, and accelerate the identification of novel drug targets. These tools have become indispensable in the era of precision oncology, where individualized treatment strategies depend on the ability to decode complex genomic, transcriptomic, proteomic, and epigenomic information.

Despite these transformative advances, several challenges remain. Validation of computational predictions in experimental and clinical settings is critical to ensure reliability and reproducibility. Interpretability of machine learning models continues to be a barrier, as many algorithms function as “black boxes,” limiting their acceptance in clinical decision-making. Furthermore, the integration of heterogeneous data sources—ranging from bulk sequencing to single-cell analyses and clinical records—requires robust frameworks to manage variability, noise, and scalability. Addressing these issues will be essential to fully realize the potential of computational oncology.

Emerging trends offer promising solutions. Explainable AI is advancing toward models that not only predict outcomes but also provide mechanistic insights, thereby bridging the gap between computational inference and biological understanding. The adoption of cloud-based infrastructure facilitates large-scale data storage, sharing, and collaborative analysis, democratizing access to computational resources across institutions and geographies. Meanwhile, single-cell analytics and spatial transcriptomics are revolutionizing cancer biology by revealing cellular heterogeneity, tumor microenvironment dynamics, and evolutionary trajectories at unprecedented resolution.

Ultimately, the success of computational oncology hinges on interdisciplinary collaboration. Computational scientists, biologists, and clinicians must work synergistically to translate in silico insights into tangible clinical benefits. This requires not only technical innovation but also the establishment of shared standards, ethical frameworks, and translational pipelines that connect computational discoveries with patient care. By fostering such collaboration, computational approaches will continue to evolve from supportive tools into central pillars of cancer research and treatment, driving the next generation of precision medicine.

Author Contributions: Conceptualisation, H.M.B.; software, H.M.B.; investigation, H.M.B.;  writing—original draft preparation, H.M.B.; writing—review and editing, H.M.B.; visualisation, H.M.B.; supervision, H.M.B.; project administration, H.M.B. The author has read and agreed to the published version of the manuscript.

Funding: Not applicable.

Acknowledgments: We are grateful to the Department of Biochemistry, College of Science, University of Jeddah, Jeddah-21589 Saudi Arabia for providing us all the facilities to carry out the entire work.

Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.

Data Availability Statement: All the related data are supplied in this work or have been referenced properly.

References

1. Vogelstein B, Papadopoulos N, Velculescu VE, Zhou S, Diaz LA Jr, Kinzler KW. Cancer genome landscapes. Science. 2013;339(6127):1546–58.
2. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646–74.
3. Kourou K, Exarchos TP, Exarchos KP, Karamouzis MV, Fotiadis DI. Machine learning applications in cancer prognosis and prediction. Comput Struct Biotechnol J. 2015;13:8–17.
4. Cruz JA, Wishart DS. Applications of machine learning in cancer prediction and prognosis. Cancer Informatics. 2006;2:59–77.
5. Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, Thrun S. Dermatologist-level classification of skin cancer with deep neural networks. Nature. 2017;542(7639):115–8.
6. Libbrecht MW, Noble WS. Machine learning applications in genetics and genomics. Nat Rev Genet. 2015;16(6):321–32.
7. Ching T, Himmelstein DS, Beaulieu-Jones BK, et al. Opportunities and obstacles for deep learning in biology and medicine. J R Soc Interface. 2018;15(141):20170387.
8. Caruana R, Lou Y, Gehrke J, Koch P, Sturm M, Elhadad N. Intelligible models for healthcare: Predicting pneumonia risk and hospital 30-day readmission. KDD. 2015;1721–30.
9. Barabási AL, Gulbahce N, Loscalzo J. Network medicine: a network-based approach to human disease. Nat Rev Genet. 2011;12(1):56–68.
10. Han J-DJ, Bertin N, Hao T, et al. Evidence for dynamically organized modularity in the yeast protein–protein interaction network. Nature. 2004;430(6995):88–93.
11. Yildirim MA, Goh KI, Cusick ME, Barabási AL, Vidal M. Drug-target network. Nat Biotechnol. 2007;25:1119–26.
12. Guney E, Menche J, Vidal M, Barabási AL. Network-based in silico drug efficacy screening. Nat Commun. 2016;7:10331.
13. Beerenwinkel N, Schwarz RF, Gerstung M, Markowetz F. Cancer evolution: Mathematical models and computational inference. Syst Biol. 2015;64(1):E1–E25.
14. Kitchen DB, Decornez H, Furr JR, Bajorath J. Docking and scoring in virtual screening for drug discovery: Methods and applications. Nat Rev Drug Discov. 2004;3(11):935–49.
15. Meng XY, Zhang HX, Mezei M, Cui M. Molecular docking: A powerful approach for structure-based drug discovery. Curr Comput Aided Drug Des. 2011;7(2):146–57.
16. Hollingsworth SA, Dror RO. Molecular dynamics simulation for all. Neuron. 2018;99(6):1129–43.
17. Pushpakom S, Iorio F, Eyers PA, et al. Drug repurposing: Progress, challenges and recommendations. Nat Rev Drug Discov. 2019;18(1):41–58.
18. Hasin Y, Seldin M, Lusis A. Multi-omics approaches to disease. Genome Biol. 2017;18(1):83.
19. Ritchie MD, Holzinger ER, Li R, Pendergrass SA, Kim D. Methods of integrating data to uncover genotype–phenotype interactions. Nat Rev Genet. 2015;16(2):85–97.
20. Meng C, Zeleznik OA, Thallinger GG, et al. Dimension reduction techniques for the integrative analysis of multi-omics data. Brief Bioinform. 2016;17(4):628–41.
21. Wang B, Mezlini AM, Demir F, et al. Similarity network fusion for aggregating data types on a genomic scale. Nat Methods. 2014;11(3):333–7.
22. Hopkins AL. Network pharmacology: The next paradigm in drug discovery. Nat Chem Biol. 2008;4(11):682–90.
23. Lee J-S, Das A, Jerby-Arnon L, et al. Harnessing synthetic lethality to predict response to cancer treatment. Nat Commun. 2018;9(1):2546.
24. Litjens G, Kooi T, Bejnordi BE, et al. A survey on deep learning in medical image analysis. Med Image Anal. 2017;42:60–88.
25. Lambin P, Leijenaar RTH, Deist TM, et al. Radiomics: The bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol. 2017;14(12):749–62.
26. Grossman RL, Heath AP, Ferretti V, et al. Toward a shared vision for cancer genomic data. N Engl J Med. 2016;375(12):1109–12.
27. Boettiger C. An introduction to Docker for reproducible research. SIGOPS Oper Syst Rev. 2015;49(1):71–9.
28. Price ND, Magis AT, Earls JC, et al. A wellness study of 108 individuals using personal, dense, dynamic data clouds. Nat Biotechnol. 2017;35(8):747–56.
29. Doshi-Velez F, Kim B. Towards a rigorous science of interpretable machine learning. arXiv. 2017;arXiv:1702.08608.
30. Piwek L, Ellis DA, Andrews S, Joinson A. The rise of consumer health wearables: Promises and barriers. PLoS Med. 2016;13(2):e1001953.
31. Svensson V, Teichmann SA, Stegle O. SpatialDE: Identification of spatially variable genes. Nat Methods. 2018;15(5):343–6.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of Global Journal of Basic Science and/or the editor(s). Global Journal of Basic Science and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. 

Copyright: © 2025 by the authors. Submitted for possible open access publication under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).