Target discovery is one of main activities in early drug development. Although a common activity, it raises many difficult challenges including: 

• There are potentially thousands of targets to consider; how do you compare all these options and select the best ones? 
• How do you avoid literature bias to find promising targets that have good underlying biological data but little mention in literature? 
• How do you enable your domain experts to evaluate a potential target, especially in new classes of targets where you may lack prior experience and expertise? 

Euretos addresses these challenges through a unique combination of novel predictions from computational disease models coupled with the ability for domain experts to assess these independently in the Euretos AI platform.

Based on its many collaborations with leading pharma and biotech companies on target discovery, Euretos has developed unique, AI-powered computational disease models that predict truly novel target candidates for a disorder by:

• Focusing on genetic associations and biological networks data, minimising potential literature bias
• Taking into account target interaction networks to assess yet unexplored disease associations
• The ability to integrate proprietary target-specific data in order to obtain the most relevant predictions

We know from many previous projects that the ability to assess predicted targets by domain experts is crucial to move a candidate forward in the discovery process. We therefore empower researchers to independently evaluate predicted targets through:

• An AI-integrated knowledge base which harmonises all supporting evidence from biomolecular databases, publications and patents
• Easy-to-use applications to evaluate all relevant aspects of target-indication involvement
• A collaborative research environment which connects disease experts and Euretos research consultants. 

Having proven computational disease models and the target assessment environment already in place, we can significantly increase the discovery potential and success of target discovery and assessment.

Euretos has created computational disease models that have been tried and tested in many collaborations with leading pharma and biotech companies.

We predict novel targets by assessing whether a target’s biological interactors are linked to known disease associations. In particular we assess:

• Target interactions in various types of biological molecular networks such as  expression correlation, protein – protein interactions and target perturbations.
• Disease - gene associations in particular genetic associations from sources like Clinvar (genomic variation) and genome-wide association study (GWAS).

For each target, we use machine learning to evaluate for thousands of indications the potential associations the gene’s multiple biological networks may have with each disease. We combine the weighted predictions from these multiple algorithms into a single gene-disease score.

The end result of this is a ranked list of targets purely based on biological data and known disease indications, minimising potential literature bias.

The ability to assess the AI-predicted targets by domain experts is crucial to move a candidate forward in the discovery process. For nearly 10 years, Euretos has been developing the Euretos AI Platform which offers a one-of-a-kind biomolecular context for target exploration & assessment.

The Euretos knowledge base uses AI technologies, such as machine reading, to harmonise knowledge from biomolecular databases, publications and patents. This includes biological interaction networks, gene expression in tissues, cell types, subcellular location and tumors, pathway interactions, regulatory and signaling responses, clinical variant associations to disease and phenotypes, clinical trial outcomes, target tractability and patents.

The Euretos AI Platform consists of multiple applications that empower researchers to draw their own, evidence-based conclusions providing them with:

• Integrated search of literature, databases and patents
• Search automation using sets for massive Pubmed searches and references heatmaps
• Detailed provenance control and biological filtering of results
• Systematic review based on key biomolecular categories (genes, pathways etc)
• Target assessment workflows, including target PPI network analysis
• Set enrichment and ranking analysis of gene interaction networks identified in the computational disease model (e.g. expression correlation, PPI and target perturbation)
• Uploading of proprietary data (such as expression data) for set analysis
• Target interaction analysis based on biological network visualisation in a relation map

The unique combination of novel predictions from the Euretos computational disease models with the ability for domain experts to assess these independently in the Euretos AI platform is very effective. Pharma and biotech companies are able to build within a very short time frame the confidence necessary to decide which targets to move forward for further development.