Computational Oncology for Precision Medicine: Biomarkers, Risk Models, and Translational Genomic Pipelines

Cancer research has rapidly entered a data-rich era. Today, researchers can draw on whole-genome sequencing, single-cell atlases, longitudinal liquid biopsy data, and increasingly detailed clinical records collected across institutions. While this expansion has opened new possibilities, it has also created a familiar challenge: we now generate far more data than we can consistently translate into clinical insight.Despite major technological advances, the impact on routine patient care has been more incremental than anticipated. The bottleneck is no longer data generation, but the ability to convert complex, high-dimensional data into robust, clinically actionable knowledge. This gap highlights a central issue in modern oncology: how to bridge research innovation with real-world clinical utility.Computational oncology sits at this interface. The field has produced sophisticated tools, including machine learning models, multi-omics integration strategies, and scalable genomic pipelines. However, their performance often drops when applied outside controlled research settings. Models trained on curated datasets may not generalize to heterogeneous clinical populations, risk prediction tools can lack reproducibility across cohorts, and analytical pipelines frequently require substantial adaptation before integration into clinical workflows.Addressing these challenges is essential if computational approaches are to meaningfully support precision medicine rather than remain largely research-focused.This Research Topic aims to bring together contributions that directly address the gap between methodological development and clinical application. We welcome submissions from computational biologists, clinicians, statisticians, data scientists, and interdisciplinary teams, spanning all career stages.We take an intentionally broad view of the field. Submissions may include new computational methods, cross-cohort validation studies, critical reviews, or well-curated and openly available datasets. The unifying goal is to improve the reliability, reproducibility, and clinical relevance of computational oncology approaches.We particularly encourage work that rigorously evaluates performance beyond initial training datasets, explicitly discusses limitations, and considers variability across patient populations and clinical contexts.Topics of Interest Areas of interest include, but are not limited to:• Biomarker discovery and validation, with emphasis on independent validation and clinical relevance• Risk prediction and patient stratification, including statistical and machine learning models integrating genomic and clinical data• Variant interpretation, focusing on distinguishing driver and passenger mutations and linking genomic alterations to therapeutic decisions• Multi-omic integration, combining genomic, epigenomic, proteomic, imaging, and clinical data, alongside frameworks to assess added value• Tumor evolution and resistance, including clonal dynamics and longitudinal monitoring approaches• Single-cell and spatial transcriptomics, particularly studies illuminating tumor heterogeneity and microenvironmental interactions• Liquid biopsy and minimal residual disease (MRD), including ctDNA-based detection, monitoring, and relapse prediction• Genomic pipelines and clinical reporting, covering end-to-end workflows, quality control, annotation strategies, and decision-support tools• Fairness, interpretability, and deployment, including evaluation across diverse populations and integration into clinical practice• Benchmarking and open resources, such as standardized datasets, reproducible workflows, and community-accessible tools Topic editors: Bharati Mehani, Hernando Lopez Bertoni, Dr Yogesh Kumar Source: https://www.frontiersin.org/research-topics/81020/computational-oncology-for-precision-medicine-biomarkers-risk-models-and-translational-genomic-pipelines