Cellular biological function, both normal and disease, emerges due to the interaction between multiple molecules in networks. Modelling and analysis of these networks provides a novel approach to drug discovery that explicitly considers the true complexity of biology. We can more effectively identify multiple prospective drug candidates with the potential to address complex disease.
We do this in a novel way: we draw from our comprehensive AI-enhanced library of millions of compounds and then use sophisticated mathematical and big data analysis techniques to match drug-like small molecules to our networks based on their potential bioactivity. These conditions, as near as possible, reflect the biological systems we are seeking to disrupt, thus increasing the likelihood of finding effective compounds.
Read more about the science behind network biology here.
Our NDD platform is a combination of large-scale, proprietary databases and a suite of powerful computational tools that employ network analysis, data mining, machine learning, AI and optimisation. The productivity increases afforded by using powerful computers at this phase of the discovery process leads to both cost and time savings. These savings facilitate the use of complex phenotypic assays, potentially closer matches to human disease biology, not traditionally suited to screening. The novelty of the approach, driven by integration of multiple disparate data sets, enables the tracking of intractable or undruggable disease processes.
See a list of our assets here
Our step by step approach is shown in the diagram below. Click on each icon below to read more about each step of our process.
Our NDD approach is driven by biology and is ideally suited for tackling many of the complex, multifactorial diseases where the needs for effective treatments remain unmet.
Our aim is to alter specific functions that are driving disease. We design interventions by selecting which mechanisms to perturb rather than by selecting in advance which targets to drug. By taking this 'bottom up' approach, NDD can find drugs that act through known or novel targets.
We explicitly model the complex cellular mechanisms involved in the disease processes we are aiming to disrupt. Network construction aims to uncover and address the redundant and degenerate pathways and sub-networks that can be missed by other approaches. Data driven network construction approaches can identify and address novel molecular mechanisms involved in disease.
Our core approach utilises biological network analysis using multiple data sets and computational tools - not just AI. Biological networks are robust by their very nature and hard to disturb. Our proprietary network analytics aim to identify molecular perturbation patterns that can significantly impact those networks and thus the disease mechanisms they represent.
The impact of millions of individual compounds on network integrity is assessed using their biological footprint of both direct and indirect protein modulations. These footprints are constructed via a statistical integration of machine learning based predictions with empirical evidence, from both structured databases and advanced natural language processing. Compounds with a high impact relative to random are selected for screening. Our in silico output are lists of compounds statistically enriched in actives.
Our hypothesis-based approach generates compound deck sizes ideal for screening in complex phenotypic screens which are more representative of human disease processes. Multiple projects, across diverse areas of biology, have demonstrated high success rates in identifying compounds with significant activity in multiple cell-based assays. These active 'hit' compounds are rational starting points for medicinal chemistry optimisation.
Active compounds across multiple chemotypes are progressed into medicinal chemistry optimising multiple properties including efficacy, DMPK and chemical novelty. Multiple projects have now optimised identified hits into leads and composition of matter patents have been filed in two programmes.