Abstract

Medicines Discovery is hard. With high failure rates, often
late in the discovery pipeline and ever-increasing costs it is critical that we
look for new ways to innovate in this field. Here at the Medicines Discovery
Catapult, we seek to do just that using a data driven, patient centric
approach. We have a highly inter-disciplinary and collaborative team with
expertise ranging from protein informatics to systems biology; imaging to
genomics, and cheminformatics to data science and software engineering; we are
here to help you innovate and are on the lookout for those “someone really
needs to” challenges!​

Data are being generated almost faster than their impact can
be understood but are absolutely critical to our understanding and ability to
treat patients. From high throughput screening to highly specialised wet-lab
experiments, data has a central place in our ecosystem. ​

We have utilised Natural Language Processing (NLP)
techniques combined with text mining to identify new elements for knowledge
discovery and developed novel models with our SME partners. Knowledge extraction
from clinical trials data combined with NLP has led to successful
collaborations with patient-led charities in the area of drug repurposing and
production of target product profile. We use a combination of chem- and
bioinformatics techniques to assess the druggability of targets and perform
high-throughput screening to help partners drive forward their target and small
molecule assets. Combining our state-of-the-art platforms for multiplex tissue
imaging and medical imaging technology, informatics is embedded into our
biomedical research infrastructure to advance the understanding of biological
samples to a range of clinical subjects. Additionally, workflows encompassing
automated identification and extraction of relevant data from graphs is applied
to several projects with great success, as have systems biology and knowledge graph-based approaches.​