Streamlining Target Identification: A Diabetes Use Case

Type 1 diabetes (T1D), characterized by the inability to produce insulin, is an autoimmune disease with no cure. Management strategies require daily insulin administration, monitoring of blood sugar levels and lifestyle changes. With >13 million cases expected by 2040,¹ there is an urgent need for innovative treatments. Target identification is a key step in this process.

Target identification using traditional keyword searching is time-consuming. With 100,000+ publications on T1D in PubMed, users are subject to reading fatigue and bias, with most readers only reading the title and abstract.² By separating the signal from the noise, Causaly can reduce reading time from months to minutes, increasing research productivity by up to 90%.

Causaly uncovered around 2,000 targets for T1D. 1,400+ targets were reported in the last 5 years. Targets were prioritized by primary data, uncovering 800+ potential targets. Using Causaly’s filtering capabilities, targets can easily by prioritized by target class. The most dominant categories for T1D targets were enzymes, signaling proteins and receptors, as shown in Figure 1.

Owing to the role of signaling pathways in promoting β-cell regeneration and modulating immune response, we have zeroed in on signaling protein targets for T1D.  Almost 120 signaling proteins have been implicated as T1D targets in primary data since 2018, according to Causaly.

Around half of these signaling proteins have been studied in animal models, the majority of which have been studied in mice. In this use case, targets were selected based on evidence strength, uncovering Fibroblast Growth Factor-21 (FGF21), suppressor of cytokine signaling 2 (SOCS2) and CXC motif chemokine ligand 12 (CXCL12) as potentially interesting signaling proteins implicated in T1D.

The rising prevalence of T1D demands new treatments. The identification and understanding of targets for T1D for developing effective drugs and alleviating patient burdens. AI can streamline this process by dramatically reducing reading time, minimizing bias and uncovering hidden target-disease insights, enabling the exploration of more promising avenues.