What are the current hot topics in the study of critical points during the process of cell mutation and development into tumors? How to conduct in-depth research on the future use of AI?

Tumorigenesis is not a linear accumulation of mutations but rather a dynamic process punctuated by critical transition points—windows during which normal cells cross irreversible thresholds toward dysplasia, carcinoma in situ, and ultimately invasive malignancy. Understanding the molecular and ecological drivers of these “tipping points” is fundamental to early interception and prevention. I would like to ask the cancer biology and computational oncology community: Which specific transition stages are currently receiving the most intense investigation—for example, the switch from field cancerization to clonal expansion, immune evasion checkpoints, metabolic reprogramming gates, or senescence escape mechanisms? What experimental models (organoids, lineage tracing, in vivo barcoding) are proving most informative? More importantly, how should AI be deployed to go beyond conventional omics pattern recognition? Are we moving toward AI systems capable of inferring causal trajectories from multi-modal longitudinal data, predicting individual-specific transition risks, or identifying hidden attractor states in epigenetic landscapes? I am also keen to hear critical perspectives on current limitations—what can AI not yet do in this domain, and where are the gaps in training data or biological theory? Insights from systems biology, single-cell genomics, and AI-driven drug discovery are all warmly welcomed. Thank you.

Post an Answer

Sign In to Answer
Accepted
0
Leslie Recio
The hottest topic currently in mutagenicity assessments is the development of tools that enable the direct quantification of mutations in DNA (Yauk et al., 2026). This technology referred to as error corrected next generation sequencing (ecNGS) enables the direct detection and quantification from a DNA sample of the 1 in 106 mutations in DNA and as importantly determination of the mutational spectrum (fingerprints) in DNA - the distinct pattern that arises from various mutational processes that occur during carcinogenesis. ecNGS enables direct interrogation in DNA of mutations in DNA that arise by either endogenous (e.g., clonal expansion of pre-existing mutant cells) and exogenous processes that can be directly integrated into wild-type rodent regulatory toxicology studies (does not require TGr animals) and going forward in human relevant NAMs. This would mean reduced reliance on hazard i.d. bioassays in DNA repair defective bacteria/p53 compromised rodent cells combined with irrelevant highly induced CYP450 bioactivation without Phase II metabolism not useful for quantitative risk assessment. As metabolically competent human relevant NAMs are developed, ecNGS can be used to directly assess the dose-response of potential mutagenic concerns and integration with IVIVE will enable quantitative risk assessments of potential mutagenic hazards. ecNGS is on track for a potential OECD test guidance application in 2027 as an alternative to TGr rodents.

The second hottest topic is the application of transcriptomic biomarkers to assess genotoxicity from transcriptome profiling in human and rodent cells  (TGxDDI/GENOMARK) (Froetschl et al., 2025). These biomarkers are highly predictive of genotoxicity in human cells and can be used retrospectively in human relevant models with pre-existing transcriptomic data and current data sets using an open source classification tool. The TGx DDI biomarker has been integrated into the US EPA NAMs program and is near acceptance by FDA as a biomarker of genotoxicity in human cells.

These new tools, TGx DDI/GENOMARK ecNGS, that enable the assessment of genotoxicity and mutagenicity in human relevant metabolically competent NAMs will enable the movement of genetic toxicology test systems from hazard i.d. to biological test systems that can be used for human risk assessments. This will be a paradigm shift in genetic toxicology assessments enabling the direct assessment of potential genotoxicity as toxicology shifts away rodent based test systems to human relevant NAMs.