Innovation
We believe that discovery research is the engine that drives innovation. SBI leads discovery research by combining computational and mathematical analysis with state of the art biological and biomedical experimentation to produce innovations of public and commercial interest.
Kinase Inhibitors: New ways to break resistance
Challenge: RAS oncogenes drive ~30% of human cancers including prevalent and deadly forms, such as colorectal, lung, pancreatic cancers and malignant melanoma. Although RAF kinases are key effectors of oncogenic RAS signaling, RAS driven cancers are usually refractory to RAF kinase inhibitors due to the formation of drug resistant RAF dimers.
What we offered:
- We have developed a next generation computational model (NGDM) that accurately predicts combinations of RAF inhibitors that block RAS signaling and transformation. The NGDM considers genetic background, network context, posttranslational modifications and thermodynamic properties of RAF dimers to identify effective drug combinations. Kinase dimerization is a general mechanism of drug resistance, and initial work indicates that our NGDM also can overcome resistance to other kinase inhibitors, such as ERBB and JAK inhibitors.
Activity/Output:
- Patent filed
- Clinical trial (ongoing)
Publication in Cell Systems - (opens in a new window)Dissecting RAF Inhibitor Resistance by Structure-based Modeling Reveals Ways to Overcome Oncogenic RAS Signaling
Contact: (opens in a new window)Oleksii Rukhlenko, (opens in a new window)Boris Kholodenko
OncoNamics: Novel Chemotherapy Predictive Diagnostic Test
Challenge: Neuroblastoma is one of the most common solid tumours in children. All high-risk neuroblastoma patients receive chemotherapy, which can have severe side-effects. Unfortunately, this chemotherapy is ineffective in 1/3 of cases. No current diagnostic tool can predict which patients will respond to and benefit from chemotherapy.
What we offered:
- A chemopredictive diagnostic test based on computational models of individual patients. The tool can predict whether chemotherapy will work for a neuroblastoma patient or not. The test will help oncologists to make better, more informed treatment decisions that avoid side effects and maximize therapeutic benefits.
Activity/Output:
- Enterprise Ireland commercial case feasibility support grant
Publication in Science Signaling - (opens in a new window)Signaling pathway models as biomarkers: Patient-specific simulations of JNK activity predict the survival of neuroblastoma patients
Patent filed - (opens in a new window)Wo2020089202 - Method for predicting the effectiveness of treatments for cancer patients
Contact: (opens in a new window)Walter Kolch
TOMOE: A discovery informatics platform to predict kinase – substrate phosphorylation networks
Challenge: Phosphorylation networks are altered in all serious diseases, including cancer. These phosphorylation networks consist of protein kinase that regulate other proteins, the substrates, by attaching a phosphate group. Kinase inhibitors have become a major new class of drugs for the treatment of cancer and inflammatory diseases. There are tens of thousands of such kinase-substrate pairs that could be useful targets for drug development, but we only know a fraction.
What we offered:
-
We have developed LinkPhinder, a software for the discovery of kinase-substrate pairs that uses the latest AI technology in knowledge graphs. LinkPhinder outperforms existing tools. It almost doubles the number of kinases for which we can identify substrates, and its predictions achieve >96% specificity and sensitivity. Thus, the accurate and comprehensive identification of kinase-substrate pairs becomes a reality.
Note: Work carried out in collaboration with Fujitsu Laboratories Ltd., the Insight Centre for DataAnalytics
Activity/Output:
- Collaboration Research Agreement
Publication in PLOS Computational Biology - (opens in a new window)Accurate prediction of kinase-substrate networks using knowledge graphs
Contact: (opens in a new window)Walter Kolch
LabPortal
Challenge: Managing all aspects of a modern biomedical laboratory is a big challenge for smaller biotech and academic laboratories that cannot afford or need the highly complex and expensive Laboratory Information Management Systems (LIMS) available for industrial laboratories. Many organizations wanting to install or upgrade a LIMS struggle to find a solution that is fit for purpose.
What we offered:
- We have developed a thin-client, web-based LIMS, which takes care of all routine tasks in the modern biomedical laboratory, including sample management, workflow management, inventory management, ordering and purchasing management, health and safety management and reporting. Being developed by researchers for researchers, it has an intuitive user interface, a self-explanatory structure, and expandable flexibility for including new tasks as they arise.
Activity/Output:
- Demo (on trial)
- Invention disclosure form
Contact: (opens in a new window)Philip Cotter, (opens in a new window)Amaya Garcia Munoz