No abstract

Genomic profiling has shown that not all cancer patients who share similar macro- and microscopical features harbour the same underlying molecular mechanism. This suggests the urge for matching patients to mechanism-based cancer therapies, independent of their primary tumour location and histology (Bashraheel et al. 2020). Currently, precision oncology trials provide personalised treatments based on the druggable variants found in a patient genetic makeup. Typically, those trials target single genetic variants (Schmidt et al. 2016; Murciano-Goroff et al. 2020) or provide combination therapies targeting single, mechanistically unrelated proteins which have been proven to be ineffective and or insufficient (Choobdar et al. 2019; Lazareva et al. 2021). In parallel, these variants are allocated to so-called canonical signalling pathways, e.g., KEGG pathways, Wiki Pathways. However, these are rather curated mind maps only combining similar signalling proteins or messengers. They do not represent true cellular signalling entities. Alternatively, signalling modules can be constructed in an unbiased manner from the interactome using validated seed proteins, also termed cancer driver genes, resulting in fragments and often mixtures of the above curated pathways (Nogales et al. 2022). These modules likely represent the true cancer mechanism and concerted network modulation with multiple mechanistically related drugs all acting on the same module i.e., network pharmacology, promise to be much more effective than targeting single unrelated variants (Cheng et al. 2019). As complex tumours will require multiple drugs targeting several modules (Sanchez-Vega et al. 2018), we start with low complexity tumours with a low mutational burden, e.g., thyroid cancer and diffuse intrinsic pontine gliomas (DIPG) (Vogelstein et al. 2013). Here, we (i) construct de-novo disease modules to identify drug targets and repurposable drugs, (ii) apply diagnostic assays to detect the patient-specific perturbed modules and (iii) decide on the therapeutic strategy to correct the modules using network pharmacology. Repurposable drugs are ranked based on clinical feasibility and other parameters. This allows a fundamentally new approach to cancer therapy often using low-side effect drugs acting in concert to improve patient survival and quality of life by implementing biology-informed drug interventions.

Noncommunicable diseases (NCDs) have become globally abundant, yet the therapeutics we use for them are imprecise. In parallel, identifying new treatments has become more costly than ever due to the ever-aggravating efficacy crisis drug discovery faces. What unites these failures is our ontological classification of diseases, primarily based on descriptive terms. To achieve precision diagnosis and precision therapy in clinical practice, NCDs need to be redefined and subdivided based on their causal molecular mechanisms. However, the inconsistency and incompatibility of the current disease classification systems hinder data integration and analysis towards the characterization of such mechanisms. Here, we explain flaws in the current disease definitions and the dispersion among existing ontologies with the aim of establishing a mechanism-based classification of diseases hence, precision medicine.

Cancer is a heterogeneous disease characterized by unregulated cell growth and promoted by mutations in cancerdriver genes some of which encode suitable drug targets. Since the distinct set of cancer driver genes can varybetween and within cancer types, evidence-based selection of drugs is crucial for targeted therapy following theprecision medicine paradigm. However, many putative cancer driver genes cannot be targeted directly, suggestingan indirect approach that considers alternative functionally related targets in the gene interaction network [1]. Oncepotential drug targets have been identified, it is essential to consider all available drugs. Since tools that offer supportfor systematic discovery of drug repurposing candidates in oncology are lacking, we developed CADDIE, a webapplication integrating six human gene-gene and four drug-gene interaction databases, information regardingcancer driver genes, cancer-type specific mutation frequencies, gene expression information, genetically relateddiseases, and anticancer drugs. CADDIE offers access to various network algorithms for identifying drug targetsand drug repurposing candidates. It guides users from the selection of seed genes to the identification of therapeutictargets or drug candidates, making network medicine algorithms accessible for clinical research. CADDIE isavailable at https://exbio.wzw.tum.de/caddie/ and programmatically via a python package athttps://pypi.org/project/caddiepy/.