Accelerate Target Discovery with the Help of AI

Drug discovery is complex, with a high level of uncertainty full of critical decisions with significant time and cost implications. With 90% of drugs failing in clinical trials,¹ there is an urgent need to accelerate the pathway from disease biology to candidate selection.

Causaly has a differentiated AI solution for preclinical research teams, empowering users to enhance productivity and make better research decisions by accelerating target discovery. Here, we identified, prioritized, and assessed potential targets for type 2 diabetes (T2D) using Causaly Cloud.

Extracting meaningful insights from 235,000+ documents on T2D in PubMed could take months, if not years. As a result, most researchers often only read the title and abstract,² introducing a bias into the papers they choose to read.

Instead, Causaly cuts through the noise, extracting only the most relevant insights, dramatically reducing reading time and bias. Using Causaly, evidence for 4,800+ targets were extracted from the literature, providing an instant overview of the T2D target landscape.

In this example, T2D targets were prioritized by data source, unveiling 1,300+ T2D targets reported in the GWAS catalog. Around 80 of these targets were reported in 2023. Here, we selected IDE as a recently reported T2D target in the GWAS catalog.

Causaly: In the GWAS catalog, rs11187007-A (which maps to IDE) was identified as a recently reported variants or risk alleles for T2D reported in a genotyping study of the China Kadoorie Biobank.³

PubMed: Performing the same search in PubMed (using keywords “targets”, “type 2 diabetes” and “GWAS” and filtering to 2023; search performed on 19/12/2023) did not return the genotyping publication identified by Causaly.

Traditional keyword searching often yields incomplete results, emphasizing the significant advantage of using human-centric AI to extract comprehensive insights.

Using Causaly Cloud, drug discovery scientists can get an instant view of all scientific documents (even those on page 1,000 of PubMed) and find hidden connections. Causaly’s high precision knowledge graph visualization also helps to enhance disease understanding and prioritize research efforts. For example, ~170 genes and proteins may mediate the effect of IDE on T2D, as shown in Figure 2.

Causaly’s AI machine-reads the entire volume of biomedical literature to extract only the most relevant insights, increasing research productivity by up to 90%. By understanding context and nuance, Causaly’s AI can uncover hidden connections from comprehensive data sources, enabling researchers to drive more effective target selection.