Computational Biology and Artificial Intelligence

Industry Collaborators

Academic Collaborators

Industry Collaborators

Academic Collaborators

Alumni
Noah Collins
Woochang Hwang
Seungbeom Lee
Méabh MacMahon
Sophia Ramani
Ming Zeng
Soorin Yim
Lilly Wollman
Mark Wong
Rebecca Marrow
Louise Lecointre
Sanjay Rathee
Anika Liu
Lena Lampe
Ivan Zyuzin
Boris Vasilev
Winnie Lei​

 Milner-LifeArc collaboration for novel disease signatures in oncology and infectious disease

The Milner Computational Research team, led by Namshik Han, have been collaborating closely with the medical research charity LifeArc over the last 3 years to develop and apply bioinformatics tools to identify novel therapeutic targets, biomarkers and drug repositioning opportunities.

By applying cutting-edge computational methods and analysis of multi-modal biomedical datasets, the team have identified new or better drug targets across a range of therapeutic areas. Key areas of research include the development and application of artificial intelligence, machine learning, statistical and mathematical approaches to pharmacogenomics and drug discovery.

Oncology research. LifeArc brings a rich oncology drug discovery programme to this collaboration. By developing methods to integrate LifeArc proprietary datasets with public cancer databases, the team have created a bespoke AI platform and computational pipeline to reveal novel disease signatures and then conducted agnostic pathway analysis to uncover new targets. Key outcomes from this work have included identification of new disease signatures for six cancer types in the LifeArc programme, and repositioning of an existing LifeArc target as a potential therapeutic strategy for an unanticipated cancer type.

COVID research. With the support of LifeArc, the group have also applied their bespoke AI methods and network analysis to uncover hidden pathways and reposition 200 already approved drugs against COVID-19. 20% of these drugs are currently in COVID-19 clinical trials, and the group went on to successfully validate two drugs in cellular assays (Proguanil and Sulfasalazine). These SARS-CoV-2 induced pathways define a resource for repurposing of drugs against COVID-19, either as monotherapies or in combination therapy.

“This collaboration has benefited enormously from the expertise of the LifeArc team in drug discovery. We see great potential to now build on the database resources and pipelines we have created together and apply them to reveal new disease signatures in oncology and other areas with high unmet medical need” says Namshik.

Dr David Pardoe, Head of Technology Development at LifeArc said:

“It is essential that we apply the power of massive and complex data to better understand fundamental biology and disease pathways. Doing this through the integration of different scientific disciplines represents just one of the ways we can collaboratively create real impact for patients”.

The Milner Computational Research Group have been working with Storm Therapeutics for the past two years to apply machine learning techniques for prediction of novel human RNA methyltransferases.

Storm Therapeutics, co-founded by our Director Tony Kouzarides and Eric Miska (Gurdon Institute) is a Cambridge-based biotech company targeting RNA-modifying enzymes, including RNA methyltransferases, in the treatment of cancer.

RNA methyltransferases are known to functionally regulate RNAs, and have attracted increasing interest as potential drug targets given their newly identified role in cancer. There are many methylation sites known to exist on RNA, but enzymes have not been ascribed for many of them, suggesting that there are a number of yet to be identified RNA methyltransferases. Georgia Tsagkogeorga, a senior scientist at Storm, set out to identify new enzymes in this family with the Milner group, using the power of machine learning. By collecting and collating transcriptomic, proteomic, structural and protein-protein interaction data for human RMTs into a harmonized database, she built a large novel machine learning dataset. She has now applied four supervised machine learning models to this database, each of which produced novel predictions of RMTs. By consolidating the model outcomes to predict novel RMT genes, the team have now selected putative novel RMTs for further validation.

“This approach is demonstrating the strong potential of machine learning to interrogate large scale databases consisting of integrated multi-omics datasets and thereby make unanticipated predictions about this enigmatic enzyme family. We are excited about the potential of this partnership to identify new epigenetic regulators with therapeutic significance” says our Director Tony Kouzarides.