In-silico Drug Discovery Service
Specialized in-silico analysis and discovery services for drug discovery research
While traditional target search methods have mainly been empirical and hypothesis-driven target searches based on individual molecules and limited data, our main difference is that we use large-scale, comprehensive data integration and AI, bioinformatics and chem informatics methods to enable more systematic and higher-dimensional target searches and biomarker searches, such as classification of disease-related targets, biological function assessment using GO-MoA, searching for new targets using pocket similarity, and verification using PPI networks.
Explore biomarkers and targets from disease
Unlike traditional methods of verifying molecules that are associated with known pathological mechanisms, such as specific proteins, metabolites, and gene mutations, one by one, it is characterized by functional exploration that combines genes related to various Biomarker extracted from PubMed papers with functional information from Gene Ontology (GO) and the mechanism of action (MoA) of compounds.
By allowing you to look at the entire disease-related molecular network through biological functions (GO terms) and expression variation patterns, you can estimate biomarkers and target groups that characterize disease from a combination of multiple molecules and pathways. This allows for more comprehensive biomarkers and target selection based on disease understanding.
An exhaustive approach using bioinformatics techniques
Explore targets from Small molecules
A method for discovering new targets is characterized by target prediction from compound structure and pocket shape similarity analysis to discover unknown targets with binding sites similar to existing targets, expanding the target network to predict related diseases.
How to discover new targets linked to disease relationships
Unlike conventional methods, many target candidates can be comprehensively extracted from compound structures by target prediction, database search, and AI analysis of literature.
Drug Repurposing
Drug repurposing is carried out in full use of target searches from diseases and target searches from compounds.
Target predictions from compound structures, literature and databases are analyzed using AI to comprehensively extract a large number of target candidates.
It expands through target predictions from structure and protein similarity, and in addition to routes to search for new disease targets and routes to explore new targets using the biological functions of diseases and their therapeutic agents, we also provide a more accurate answer by examining AI (MNLI) based on hypotheses tailored to papers and various texts.
Drug repurposing starting from diseases with AI(MNLI) technology
By classifying the relationships of disease-related compound information using AI, expanding targets based on pocket similarity from existing targets and adding PPI, drug repurposing from various fields.
In silico prediction of phenotype screening data
This service predicts functions such as toxicity phenotypes and mechanisms of action from gene expression experiments performed in phenotypic screening.
By preparing a gene set specialized for a function in advance and evaluating expression data in model animals using the GSEA method, function can be predicted.
Here, we show an example of predicting a phenotype related to developmental toxicity.
A gene set was created for use in GERA analysis to predict developmental toxicity phenotypes.
In this example, the relationship between phenotypes and genes in the field of developmental toxicology was created from PubMed and GeneRIF data using text mining and NMLI based on an ontology dictionary.
(This service creates gene sets using a pipeline that combines various databases according to the purpose.)
This is an example of analysis in zebrafish that has been extrapolated to humans for developmental toxicity. By constructing a gene set related to developmental toxicity using AI and text mining, and performing GSEA analysis using gene expression experimental data, functional analysis of phenotype screening is possible. World Fusion is able to construct various gene sets, and by using homologenes, we are able to screen and analyze specific to various phenotypes.
For zebrafish, we can also work with research institutes to provide services from assay planning.
Collaborating organizations: Department of Systems Pharmacology, Mie University Graduate School of Medicine
About GO-MoA
GO-MoA is a data resource that statistically analyzes the relationships between target genes of drugs (including investigational drugs) indicated for diseases and the biological processes defined by Gene Ontology (GO). By utilizing this data, GO-MoA clarifies which biological functions are involved in disease treatment and helps identify drugs that may act through the same biological processes. It is also useful for discovering new therapeutic target genes and evaluating the potential for drug repurposing.
By focusing on the biological processes of genes targeted by effective drugs for specific diseases, GO-MoA can suggest unexplored genes with similar functions as potential new targets. Additionally, it evaluates the possibility of repurposing existing drugs for diseases beyond their original indications, creating new opportunities for drug repurposing.