A huge thank you to the entire team and collaborators whose work made this possible! [5/5]

Our findings suggest that cytarabine could provide a targeted treatment option for a genetically defined group of patients with leptomeningeal dissemination of glioma, making an existing therapy newly actionable through biomarker guidance. [4/5]

This work highlighted a strong dependency on DNA polymerase pathways and pointed us toward cytarabine, an existing FDA-approved drug. When we tested cytarabine across several glioma models, we found that tumours with both TP53 and RB1 intact were significantly more sensitive than others. [3/5]

In this study, we used large-scale CRISPRi screening in patient-derived glioma models to identify the genetic weaknesses of these tumours. [2/5]

If you are a UK citizen or resident, please lend your voice to this cause! [3/3]

Their mission is our mission: to accelerate progress in brain cancer treatment. They're calling on the UK government to increase research funding, standardise genomic sequencing for better-informed treatment, and enact "Right to Try" laws for those with terminal diagnoses. [2/3]

#CoSyneTherapeutics #Glioblastoma #BrainCancerResearch #BioTech #KnowledgeGraphs

Instead of just relying on the data within the LLM compressed in its weights, the prompt contains now additional information from structured resources. Naturally, this can be also leveraged in agentic workflows that chain different tasks executed by the LLM together. [13/13]

🤖 Generative AI enhancement: Knowledge Graphs can provide structured data to large language models (LLM; for example ChatGPT) to reduce hallucinations and improve accuracy (known as retrieval augmented generation). [12/13]

📚 Biomedical literature: Which research groups are making similar discoveries? How strong is the evidence for a particular claim made in a research publication? Do we observe validation of findings/claims by other scientists? [11/13]

🔭 Beyond the drug repurposing use case:

Knowledge Graphs are transforming biomedical research (and other disciplines) in other ways too. Some examples: [10/13]

Instead of searching through scattered databases, we can ask complex questions like: "Show me all safety-approved drugs that target proteins we know are critical in GBM." The example below shows how this would work in practice. [9/13]

🔋 What makes Knowledge Graphs powerful?

Think of a Knowledge Graph as a network that connects related information, bridges different concepts, and shows the relationship between them. [8/13]

By mapping out these relationships, we can identify human-safe drugs that might be repurposed for treating GBM (and other deadly brain diseases). This could dramatically reduce the development time and risk typically required for the development of new treatments. [7/13]

- Which drugs have already passed safety trials?
- What proteins do they target?
- Which biological functions are these proteins involved in?
- How do those proteins and functions relate to brain cancer based on our proprietary biobank and experimental data? [6/13]