Long non-coding RNAs (lncRNAs, classically defined as transcripts >200 nt, although some sources apply a >500 nt cutoff) have emerged as essential regulators of tumor plasticity by bridging chromatin organization, metabolic homeostasis, and therapeutic outcome (Huarte, 2015). Acting as molecular scaffolds, decoys, and regulatory hubs, lncRNAs fine-tune transcriptional and metabolic networks that allow cancer cells to adapt to environmental fluctuations and pharmacological stress. Depending on the cellular context and tumor type, lncRNAs can either reinforce therapy resistance or restore sensitivity, placing them at the center of the epigenetic–metabolic circuitry that determines treatment efficacy (Grossi et al., 2025; Wu et al., 2023). A growing body of evidence demonstrates that oncogenic lncRNAs coordinate epigenetic remodeling with metabolic reprogramming to maintain proliferation, redox balance, and survival during therapy (Figure 1) (Grossi et al., 2025).

At the mechanistic level, many of the earliest and most pervasive effects of lncRNAs in cancer emerge through their ability to shape chromatin architecture and epigenetic states. By guiding chromatin-modifying complexes, organizing nuclear domains, and establishing repressive or permissive transcriptional environments (Jiang et al., 2024), lncRNAs lay the foundation upon which downstream metabolic and signaling adaptations are built. This positioning at the interface of chromatin organization and transcriptional control enables lncRNAs to act as primary regulators of cellular plasticity, making epigenetic scaffolding functions a logical starting point for understanding their role in therapy response.

Acetyl-CoA is produced by mitochondria and represents a central hub in various metabolic pathways, including the oxidation of glucose, amino acids, and fatty acids. Considering sources of acetylation, glucose is certainly one of the most important. From this perspective, given that glucose represents one of the main energy sources for many tumors, this aspect is particularly relevant in cancer. Lactate also becomes an important source of acetylation in tumors. In the case of lactate, it can be produced endogenously by tumor cells or taken up from the extracellular environment, where it is mainly generated by cells of the tumor microenvironment (Shang et al., 2022). Other important sources of acetyl-CoA include fatty acid β-oxidation and the metabolism of certain amino acids, including branched-chain amino acids (BCAAs), as well as ketone body metabolism. The contribution of acetyl-CoA derived from these different metabolic pathways depends on the cell type, the predominant metabolic program, and the cellular state (Guertin and Wellen, 2023). Beyond the biochemical role, acetyl-CoA represents a critical donor of acetyl groups for post-translational modifications, particularly histone acetylation. As such, fluctuations in acetyl-CoA availability can directly influence chromatin dynamics and gene expression, linking cellular metabolic status to epigenetic regulation and cell fate decisions (Hao et al., 2022). The role of histone acetylation in cancer is not univocal, as acetylation at different residues of various histones can lead to opposing effects on tumor behavior. This variability is also tumor-type dependent. Specifically, regarding histone H3 acetylation, certain modifications are particularly associated with specific cancer types, potentially serving as true biomarkers (Miziak et al., 2024).

Histone acetylation occurs on lysine residues of histone proteins, leading to a reduction in the positive charge of histones and, consequently, a weaker interaction with DNA. This results in a more relaxed chromatin structure, which is more accessible to the transcriptional machinery. As a consequence, genes located in these regions are more easily expressed. In general, therefore, histone acetylation is considered a modification that promotes gene expression (Shvedunova and Akhtar, 2022; Wang and Ma, 2025).

In the context of the response to chemotherapy, histone acetylation has generally been associated with the development of resistance (Figure 2). Gene ontology analysis of differentially expressed genes in cisplatin-resistant tumor cell lines compared to their parental counterparts highlighted the involvement of pathways related to epigenetic modifications. In particular, gene expression alterations observed in cisplatin-resistant HeLa and HepG2 cell lines were evident both at the level of enzymes responsible for epigenetic modifications and in histone acetylation itself. Indeed, histone acetylation appeared to be a crucial factor in the development of resistance to cisplatin (Natu et al., 2024).

Consistently, several histone acetyltransferase (HAT) inhibitors have shown promising effects in targeting cancer stem cells and overcoming drug resistance. For example, pharmacological inhibition of KAT6A in ovarian cancer models has been associated with increased cisplatin sensitivity, suggesting a potential role for HAT inhibitors in combination strategies aimed at overcoming chemoresistance (Liu et al., 2021). In addition, in castration-resistant prostate cancer, in which high expression of histone acetyltransferase 1 (HAT1) can enhance the resistance to enzalutamide and gemcitabine (Wang et al., 2023), inhibitors of histone acetylation have been proposed as therapeutic agents or adjuvant treatments, showing potential in reducing tumor burden. FLIM-FRET screening has identified a panel of 11 potential epigenetic biomarkers for breast cancer therapy. Among these, several were associated with DNA modifications, including methylation and acetylation patterns. Based on these findings, anacardic acid, a natural compound, was subsequently identified as a potential adjuvant to standard Tamoxifen therapy. Its co-administration was found to enhance Tamoxifen efficacy and reduce the likelihood of treatment resistance. Anacardic acid is proposed to exert its effects by inhibiting p300 HAT activity, leading to a reduction in H4K12 and H3K27 acetylation levels and a consequent decrease in occupancy at the transcription start site of estrogen response element (ERE)-regulated genes (Liu et al., 2019). However, for the sake of completeness, it is important to note that HDAC (histone deacetylase) inhibitors may also exert a beneficial effect by enhancing the response to therapy in castration-resistant prostate cancer. In this context, the improved efficacy of chemotherapeutic regimens combined with HDAC inhibitors is largely attributed to their impact on non-histone protein acetylation. Notably, several histone deacetylase (HDAC) inhibitors, including Romidepsin, Vorinostat, and Panobinostat, are currently in clinical development and have progressed to phase II clinical trials (Biersack et al., 2022). A phase I clinical trial is going on to evaluate the safety and maximum tolerated dose of EP31670, a dual inhibitor of BET and CBP/P300, in advanced solid tumors (Wang et al., 2023). This evidence highlights the critical role that intracellular acetylation balance may play in modulating therapeutic outcomes.

Non-histone acetylation occurs on proteins located in the cytosol or other cellular compartments, and plays a key role in regulating the activity and function of these proteins (Wang and Ma, 2025). A prominent example is p53, which can be acetylated by the acetyltransferase p300. This post-translational modification (PTMs) regulates p53 activity; for instance, increased levels of acetylated p53 in hepatocellular carcinoma are associated with a metabolic shift toward oxidative metabolism, resulting in reduced cell proliferation (Di Leo et al., 2019).

Lysine acetyltransferases (KATs) and lysine deacetylases (KDACs) play a central role in the dynamic regulation of protein acetylation by mediating the addition and removal of acetyl groups on both histone and non-histone proteins. KDACs are broadly categorized into two classes based on their cofactor dependency: zinc-dependent histone deacetylases and NAD⁺-dependent sirtuins (Li Z. et al., 2025). From a therapeutic perspective, the inhibition of non-histone acetylation can be achieved through the use of HDAC inhibitors, which, as previously mentioned, are frequently employed in combination with chemotherapeutic agents in cancer treatment. For instance, metformin has been shown to enhance the anti-bladder cancer activity of Panobinostat, likely through AMPK activation and the consequent regulation of the acetylation/deacetylation balance (Okubo et al., 2019). Further supporting the role of non-histone acetylation in therapy resistance, studies have shown that in breast cancer, cisplatin normally induces acetylation of serine-arginine protein kinase 1 (SRPK1). However, in cisplatin-resistant cells, SRPK1 acetylation is diminished, promoting its phosphorylation instead. This post-translational modification alters SRPK1 function, driving the expression of anti-apoptotic splicing variants. This evidence is particularly significant, as it suggests that targeting SRPK1 acetylation may offer a novel therapeutic approach to resensitize resistant breast cancer cells to cisplatin (Wang C. et al., 2020).

Human cells express more than 100 epigenetic enzymes that are able to modify histones, DNA and transcriptional factors. Most of them require a metabolite for their activity contributing to the emergence of a complex network system. As Wang et al. accurately described, 43 HMTs use SAM as a co-substrate; 2 LSDs that use FAD; and 21 JMJD enzymes that use alpha-ketoglutarate (Kim et al., 2020). Alterations in the DNA methylome have been associated with tumorigenesis across various solid tumors. DNA methylation, which primarily involves the addition of a methyl group to the 5-carbon position of cytosine residues within cytosine-guanine (CpG) dinucleotides, is one of the earliest discovered and most extensively studied epigenetic modifications. In general, DNA methylation leads to gene expression silencing, whereas hypomethylation, particularly in the promoter regions of oncogenes, can promote their overexpression. A similar mechanism has also been linked to genes involved in the response to chemotherapy, highlighting the role of epigenetic regulation in drug sensitivity and resistance (Romero-Garcia et al., 2020; Atlante et al., 2025). Such epigenetic alterations are not always inherited but can be acquired during tumor progression. They may directly affect oncogenes or tumor suppressor genes, or act indirectly through modifications in the promoters of their regulatory elements, such as microRNAs and lncRNAs (Mangraviti and Castelli, 2025; Romero-Garcia et al., 2020). Numerous alterations have been identified at the level of CpG islands, including both hypomethylation and hypermethylation events, and they are associated with drug resistance (Figure 2). In colorectal cancer, DNA methylation profiling has been performed in cells resistant to 5-Fluorouracil (5-FU), a widely used chemotherapeutic agent whose clinical efficacy is often limited by the development of drug resistance. As a result, five genes—CCNE1, CCNDBP1, PON3, CHL1, and DDX43—have been identified and proposed as potential therapeutic targets to overcome 5-FU resistance (Baharudin et al., 2017). Dietary folate intake plays a critical role in regulating methylation metabolism in the colorectal mucosa. Folate deficiency reduces intracellular levels of SAM, leading to global DNA hypomethylation and an increased risk of colorectal cancer. In premalignant or very early stages, limiting one-carbon unit availability can transiently suppress nucleic acid synthesis, impairing the replication capacity of highly proliferative adenomatous cells and exerting a growth-inhibitory effect. However, once tumor transformation has occurred, chronic folate insufficiency compromises thymidylate synthesis efficiency and promotes uracil misincorporation into DNA. This results in increased double-strand breaks and chromosomal instability, accelerating clonal selection and the evolution of more aggressive tumor phenotypes. The folate cycle regulates the intracellular balance between SAM and S-adenosylhomocysteine (SAH), thereby controlling the activity of DNA methyltransferases and histone methyltransferases. While modest increases in SAM levels may help preserve the epigenetic integrity of tumor suppressor gene promoters, excessive methyl-donor availability in a pro-oncogenic signaling environment may stabilize chromatin configurations that favor tumor progression (Sun Y-H. et al., 2025). Thus, the use of methylation modulators has been proposed in colorectal cancer as a strategy to overcome drug resistance. Among these, inhibitors targeting HMTs and histone demethylases (HDMs) are currently under investigation for their potential to suppress colorectal cancer cell proliferation and enhance sensitivity to chemotherapy (Haynes and Manogaran, 2025).

Another example of the close relationship between methylation and drug resistance in cancer is observed in nine esophageal cancer cell lines resistant to Docetaxel, Nedaplatin, Mitomycin, and Cisplatin, hypermethylation of the MCTP1 gene—implicated in signal transduction—has been observed. This hypermethylation leads to a downregulation of gene expression (Kong et al., 2021).

As with other histone modifications, methylation can also contribute to the generation of cancer cell subpopulations. This heterogeneity is not only genetic or epigenetic but also phenotypic, and it can influence how tumor cells respond to anticancer therapies. With regard to methylation, heterogeneous methylation patterns contribute to the formation of tumor cell clones that confer heterogeneity and plasticity to the entire tumor mass, thereby promoting the development of resistant cells as well as the emergence of dormant cells. These cells do not undergo cell death following treatment but instead remain in a non-proliferative state, retaining the potential to be reactivated later under more favorable conditions (Sadida et al., 2024). Numerous studies have demonstrated that promoter hypermethylation of genes such as MLH1, MGMT, and BRCA1 can drive resistance to different anticancer agents (Romero-Garcia et al.; Sadida et al., 2024). For example, in non-small-cell lung cancer, hypermethylation of IGFBP3 is associated with reduced sensitivity to cisplatin, while hypermethylation of the MGMT gene has been shown to influence the response to temozolomide in glioblastoma (Duruisseaux and Esteller, 2018).

The dioxygenase ten-eleven translocation (TET) enzyme promotes DNA demethylation, which is often altered in cancer. These enzymes represent a bridge between the epigenetic modification of methylation and metabolism, as TET enzymes are two-oxoglutarate (also known as α-ketoglutarate)-, oxygen-, and iron-dependent dioxygenases, and they use ascorbic acid as a cofactor (Salmerón-Bárcenas et al., 2024). Loss of TET2 expression has been shown to promote resistance to tamoxifen treatment in vivo, by promoting mammary tumor development with deficient ERα expression (Kim et al., 2020). The role of TET enzymes in chemotherapy resistance is compounded by the fact that they can perform functions not strictly associated with methylation. For example, in hepatocellular carcinoma, it has been shown that DNMT3a and TET2 act coordinately to regulate cancer cell fate through both DNA methylation-dependent and -independent mechanisms, supporting drug resistance and poor prognosis. For this reason, they have been proposed as promising therapeutic targets for refractory hepatocellular carcinoma (Cheng et al., 2024).

Histone methylation influences tumor response to radiotherapy primarily through the regulation of DNA damage repair pathways, particularly homologous recombination (HR) and nonhomologous end joining (NHEJ). Specifically, methylation marks on histone residues H3K36 and H3K27me3 are critically involved in these processes (Wen et al., 2023). For instance, trimethylation of H3K36 by the methyltransferase SETD2 is essential for the activation of ATM kinase and the recruitment of 53BP1 to sites of DNA double-strand breaks (DSBs), facilitating effective DNA repair. In contrast, overexpression of the histone demethylase JMJD2A reduces H3K36 methylation, impairs HR repair efficiency, and may thereby sensitize tumor cells to radiation. On the other hand, Metnase, a methyltransferase that also targets H3K36, promotes NHEJ repair, ultimately enhancing radiation resistance (Song et al., 2025).

Given the strong connection between epigenetic modifications and the response of cancer cells to both chemotherapy and radiotherapy, numerous epigenetic modulators have been introduced into cancer treatment as adjuvants to conventional therapies (Table 2). To date, numerous epigenetic modulators have received FDA approval for use in cancer therapy. In addition to FDA-approved agents, a wide range of epigenetic modulators—particularly DNA methyltransferase inhibitors are currently being evaluated in clinical trials across various cancer types. For example, guadecitabine is undergoing phase III trials for acute myeloid leukemia (ClinicalTrials.gov ID NCT02920008), while RX-3117 is being tested in phase I/II studies for metastatic pancreatic and bladder cancers (ClinicalTrials.gov ID NCT03189914). Other DNMTis such as 5-fluoro-2′-deoxycytidine are in phase II trials for head and neck, lung, and breast cancers (ClinicalTrials.gov ID NCT00978250).

Beyond DNA methyltransferase inhibitors, other compounds with epigenetic or DNA-modifying activity, such as hydralazine, EGCG, genistein, and curcumin, are also under investigation for their potential to enhance the efficacy of conventional therapies. For instance, hydralazine is in phase III trials for ovarian cancer (ClinicalTrials.gov ID NCT00533299), EGCG is in phase II for prostate and colon cancers (ClinicalTrials.gov ID NCT02891538), and curcumin is being evaluated for colorectal and breast cancers (ClinicalTrials.gov ID NCT01042938). Some agents, like disulfiram and resveratrol, are being repurposed for oncology and studied in multiple solid tumors, including breast, lung, and colon cancers (Zhang et al., 2022).

Undoubtedly, acetylation and methylation represent the main epigenetic modifications, both histone and non-histone, on which research has primarily focused in relation to the regulation of cancer therapy response. Beyond these, however, as previously mentioned, other epigenetic modifications may contribute to establishing a strong link between cellular metabolism and tumor aggressiveness (Figure 2).

SUMOylation is a reversible post-translational modification in which small ubiquitin-like modifier proteins (SUMOs) are covalently attached to lysine residues on target proteins through an enzymatic cascade. This modification regulates multiple cellular processes, including protein stability, subcellular localization, protein–protein interactions, transcriptional activity, and the cellular response to stress (Huang et al., 2024). In hematopoietic malignancies, inhibition of SUMOylation has been shown to enhance the activity of DNA hypomethylating agents, thereby increasing drug efficacy and limiting malignant cell expansion. In particular, the combination of the SUMOylation inhibitor TAK-981 and the DNA hypomethylating agent 5-aza-2′-deoxycytidine has been investigated as a promising therapeutic strategy to improve treatment outcomes in MYC-driven hematopoietic cancers (Kroonen et al., 2023).

Lactylation consists of the addition of lactate to proteins, including histones. The lactate used for these modifications is mainly derived from glucose, linking lactylation to glycolytic metabolism. In addition, lactate can originate from the tumor microenvironment, thereby creating a functional link between the epigenetic state of tumor cells and the surrounding microenvironment (Xu et al., 2023). Histone lysine lactylation, which primarily occurs on lysine residues of H2A, H2B, H3, and H4, has frequently been found to be altered in cancer. Elevated global levels of histone lactylation have been observed across most cancer types, where this modification contributes to the stimulation of gene transcription within chromatin. Given the critical role of lactylation in cancer development and resistance to therapy, targeting lactylation pathways represents a promising avenue for new cancer treatments. While most inhibitors targeting lactylation are still at the preclinical research stage, a few have progressed to clinical trials. Notably, compounds such as AZD3965, an inhibitor of MCT1, and 2-deoxy-D-glucose (2-DG), a glycolysis inhibitor, have undergone evaluation for their safety profiles, pharmacokinetics, and maximum tolerated doses in clinical settings (Sun Y. et al., 2025).

Lysine succinylation is a recently identified post-translational modification characterized by the covalent addition of succinyl groups to lysine residues on target proteins, thereby modulating their structure, activity, and function. Succinyl-coenzyme A (succinyl-CoA) represents the primary donor of succinyl groups, and its intracellular levels are tightly regulated by multiple metabolic pathways. These include the TCA cycle, BCAA catabolism, and the β-oxidation of fatty acids. Most cellular succinylation reactions are enzymatically regulated by the coordinated activity of succinyltransferases and desuccinylases. Through the balanced action of these opposing enzymes, both histone and non-histone proteins undergo dynamic succinylation and desuccinylation (Zheng et al., 2020). In ovarian cancer, MFF succinylation promotes the maintenance of stemness and contributes to cisplatin resistance. Glyburide was used to inhibit lysine succinyltransferase activity and block MFF succinylation (Zhu et al., 2025). Histone succinylation has also been linked to chemotherapy resistance. For example, in prostate cancer, increased levels of H3K122 succinylation have been observed. Conversely, in pancreatic cancer, elevated H3K122succ levels have been associated with increased sensitivity of cancer cells to gemcitabine (Gao and Yu, 2025).

These modifications belong to a recently classified novel class of histone “non-enzymatic covalent modifications” (NECMs), which lies at the intersection between epigenetics and metabolic fitness and plays a role in cancer, particularly in the response to therapy.

In this context, the link with metabolism is direct, as these modifications are not enzymatically driven but instead depend solely on the concentration of the metabolite. The metabolites responsible for these histone modifications are generally produced in the cytoplasm through different metabolic pathways primarily aimed at ATP generation. When the concentration of specific intermediates increases, they can freely diffuse into the nucleus, where they covalently bind to histone proteins (Zheng et al., 2020) (Figure 3).

Because these modifications, as mentioned above, are associated with the aggressiveness of many tumors, several therapeutic strategies have been developed and tested to regulate NECMs. These approaches are mainly based on targeting erasers, which are responsible for removing the modifications, and readers, which recognize and interpret them (Scumaci and Zheng, 2023; Maksimovic and David, 2021).

The Cancer Genome Atlas (TCGA) has transformed cancer research by providing integrated genomic, transcriptomic, epigenomic, and proteomic profiles from over 20,000 tumors and matched normal tissues across 33 cancer types (Cancer Genome Atlas Research Network et al., 2013). The completion of the Pan-Cancer Atlas expanded these resources by uncovering molecular commonalities and lineage-specific differences across tumor types (Hoadley et al., 2018). Analyses of TCGA datasets have revealed epigenetic vulnerabilities and metabolic dependencies with therapeutic potential (Lehmann et al., 2021; Rosario et al., 2018), demonstrated how DNA methylation and histone modifications open new treatment avenues (Gnad et al., 2015; Cheng et al., 2023), and guided the development of targeted epigenetic therapies (Fan et al., 2019).

Complementing TCGA, the Encyclopedia of DNA Elements (ENCODE) provides the regulatory context with nearly 6,000 datasets spanning chromatin accessibility, transcription factor binding, histone modifications, DNA methylation, and regulatory element activity across diverse cellular states (ENCODE Project Consortium, 2012; ENCODE Project Consortium et al., 2020; Luo et al., 2020; Subramanian et al., 2020). These data are critical for designing combination therapies that leverage epigenetic drugs to sensitize tumors to conventional treatments or bypass resistance (Klemm et al., 2019).

The Gene Expression Omnibus (GEO) serves as a central repository for epigenetic drug–relevant datasets, housing thousands of studies on drug responses, biomarkers, and resistance mechanisms (Barrett et al., 2013). With over five million samples, GEO now includes curated pharmacoepigenomics signatures that enable target validation, biomarker discovery, and patient stratification (Wang et al., 2016; Wang et al., 2019).

Extending from these descriptive resources, the Cancer Dependency Map (DepMap) provides functional insights by integrating CRISPR-Cas9 screening with multi-omics profiling (Tsherniak et al., 2017). DepMap has uncovered synthetic lethal interactions involving epigenetic regulators, revealing opportunities for precision therapies (Behan et al., 2019). Recent analyses highlight lineage-specific vulnerabilities and show how epigenetic states shape therapeutic susceptibility and resistance (Ohnmacht et al., 2023).

Taken together, these resources provide raw data landscapes for epigenetics-based drug discovery. To extract actionable insights, however, computational integration tools are essential.

Weighted Gene Co-expression Network Analysis (WGCNA) identifies drug targets by constructing co-expression modules that capture coordinated epigenetic and metabolic responses to therapy (Langfelder and Horvath, 2008; Zhang et al., 2005). Applied to TCGA, WGCNA has revealed modules linked to drug sensitivity, informing network-based biomarkers for patient stratification (Pei et al., 2017).

The Algorithm for the Reconstruction of Accurate Cellular Networks (ARACNe) reconstructs transcriptional regulatory networks from multi-omics data (Margolin et al., 2006). By identifying master regulators affected by epigenetic drugs, ARACNe reveals targets, feedback loops, and resistance mechanisms, guiding combination strategies.

iClusterPlus enables integrative subtyping by combining genomic, epigenomic, and transcriptomic data (Shen et al., 2009). It has been applied to stratify patients, predict resistance, and uncover therapeutic vulnerabilities, particularly in hematologic malignancies treated with agents such as 5-azacytidine and HDAC inhibitors (Lachmann et al., 2016; Califano and Alvarez, 2017).

When applied in concert, these tools make it possible to move from descriptive datasets to functional insights, supporting the development of precision therapies (Table 3).

Systems biology now seeks to integrate molecular, cellular, and tissue processes, particularly the coupling of epigenetic regulation (DNA methylation, histone modifications, chromatin architecture) with metabolic states (enzyme expression, flux, metabolite pools). Network biology provides a natural scaffold, representing genes, ncRNAs, proteins, chromatin elements, and metabolites as nodes, and their regulatory or biochemical relationships as edges. Multi-omics datasets enable multiplex networks that explicitly encode cross-talk among molecular layers.

Artificial intelligence (AI) offers powerful tools to decode these networks. By learning compact representations of complex, nonlinear multi-omics data, AI can capture cross-layer regulatory loops and context-specific rewiring. These embeddings support prediction, classification, and mechanistic inference, expanding the toolkit for drug discovery.

These concepts become especially powerful when applied to specific biological questions, such as ncRNA regulation of metabolism or the prediction of metabolic reprogramming from chromatin states.

Network analyses are pivotal for uncovering ncRNA regulation of metabolism. Co-expression tools like WGCNA identify ncRNA–enzyme modules, while ARACNe infers direct ncRNA–gene regulatory interactions. Methods such as xMWAS and DIABLO extend these insights by correlating ncRNAs, mRNAs, and metabolites, highlighting candidate regulatory axes (Singh et al., 2019). Neural models such as mmvec further capture ncRNA–metabolite co-occurrence patterns (Morton et al., 2019). Together, these approaches reveal ncRNAs acting as metabolic “hubs,” sequestering miRNAs or modulating enzyme expression (Li M. et al., 2025). Similar frameworks can also be applied to understand how chromatin dynamics themselves encode and predict metabolic states.

Chromatin features (DNA methylation, histone modifications, accessibility) both reflect and shape metabolism, enabling multi-omics inference of metabolic phenotypes (Ge et al., 2022). MER-Net exemplifies this integration, linking epigenomic and metabolic networks with deep learning (Wang et al., 2021). Tools like iClusterPlus identify latent epigenetic–metabolic clusters (Sathyanarayanan et al., 2019), while genome-scale metabolic and regulatory network integration predicts essential metabolic genes and synthetic lethal pairs (Barrena et al., 2023). AI-driven classifiers can now predict metabolic reprogramming (e.g., glycolysis vs. oxidative metabolism) directly from chromatin states. These predictive frameworks naturally feed into drug discovery, where the goal is to identify compounds that exploit epigenetic–metabolic vulnerabilities.

Large-scale resources such as DepMap (Ghandi et al., 2019; Meyers et al., 2017), PRISM drug responses, and Connectivity Map/LINCS (Subramanian et al., 2017) form the backbone for network- and AI-driven drug repurposing. Graph-based learning applied to heterogeneous networks of genes, ncRNAs, metabolites, proteins, drugs, and diseases has successfully nominated novel drug candidates (Zhao et al., 2022). Such approaches suggest, for example, that metabolic enzyme inhibitors may be effective in tumors with specific epigenetic profiles (You et al., 2022).

Functional genomics further reveals actionable vulnerabilities: DepMap identifies synthetic lethalities between epigenetic regulators and metabolic pathways (Barrena et al., 2023), while CMap/LINCS nominates compounds that mimic or reverse epigenetic–metabolic dysregulation (Pujalte-Martin et al., 2024). AI applied to multi-omics cohorts connects mutations in histone modifiers to metabolic dependencies and suggests existing inhibitors as therapeutic options (Salvati et al., 2025). These computational screens generate prioritized compound lists for experimental testing, aiming to align epigenetic–metabolic therapies with responsive patient subgroups.