A Resource to Infer Molecular Paths Linking Cancer Mutations to Perturbation of Cell Metabolism

Abstract

Some inherited or somatically-acquired gene variants are observed significantly more frequently in the genome of cancer cells. Although many of these cannot be confidently classified as driver mutations, they may contribute to shaping a cell environment that favours cancer onset and development. Understanding how these gene variants causally affect cancer phenotypes may help developing strategies for reverting the disease phenotype. Here we focus on variants of genes whose products have the potential to modulate metabolism to support uncontrolled cell growth. Over recent months our team of expert curators has undertaken an effort to annotate in the database SIGNOR 1) metabolic pathways that are deregulated in cancer and 2) interactions connecting oncogenes and tumour suppressors to metabolic enzymes. In addition, we refined a recently developed graph analysis tool that permits users to infer causal paths leading from any human gene to modulation of metabolic pathways. The tool grounds on a human signed and directed network that connects ∼8400 biological entities such as proteins and protein complexes via causal relationships. The network, which is based on more than 30,000 published causal links, can be downloaded from the SIGNOR website. In addition, as SIGNOR stores information on drugs or other chemicals targeting the activity of many of the genes in the network, the identification of likely functional paths offers a rational framework for exploring new therapeutic strategies that revert the disease phenotype.

Curation of causal interactions mediated by genes associated to autism accelerates the understanding of gene-phenotype relationships underlying neurodevelopmental disorders

“…Embedding pathways and phenotypes into the cell causal network allows one to walk along the directional edges of the network and to estimate a causal distance between a protein and a phenotype or a pathway. To support this type of analysis, we recently implemented ProxPath , an algorithm that, given a set of proteins, estimates its regulatory impact over phenotypes and pathways annotated in SIGNOR (Iannuccelli et al , 2022) (Figure 4B). ProxPath identifies short causal paths linking two graph nodes and estimates their functional distance .…”

“…Significantly-proximal phenotypes are identified using a strategy adapted from Iannuccelli et al (2022). The pipeline consists of two steps:…”

“…The ProxPath algorithm can estimate the regulatory impact of an input list of proteins, in our case SFARI (score 1) proteins, over phenotypes annotated in SIGNOR. Significantly-proximal phenotypes are identified using a strategy adapted from Iannuccelli et al (2022). The pipeline consists of two steps: STEP1: Browsing the SIGNOR interactome to create Phenotype-proximity annotations.…”

“…We report here a curation effort, carried out over the past couple of years, aimed at increasing this coverage. In addition, to showcase the relevance of the curated dataset in dissecting the molecular mechanisms underlying neuropsychiatric diseases, we adapted our recently-developed computational strategy, here dubbed ProxPath (Perfetto et al , 2021; Iannuccelli et al , 2022). ProxPath exploits causal information annotated in SIGNOR to extend pathway and phenotype annotation in order to connect a larger fraction of autism-related proteins to a list of cellular pathways and phenotypes.…”

SIGNOR 3.0, the SIGnaling network open resource 3.0: 2022 update

“…The degree distribution of the SIGNOR graph, as that of many biological interaction networks, follows a power law with few nodes connected by many interactions while most nodes are connected by only one or few edges (Figure 1D ). A novelty of SIGNOR 3.0 is the adoption of a model that permits the integration of metabolic reactions into a causal network ( 13 ). This now allows one to explore the crosstalk between signaling and metabolism in a single connected network.…”