Increasingly it is important to take cross-disciplinary approaches to analysing biomedical data from tissue banks and biorepositories. This is because the requisite skills are broad, crossing disciplines ranging from wet lab skills dealing with biological samples through to histopathological skills with microscopes and to computing skills writing scripts and building artificial intelligence and bioinformatics models. Research advances quickly in all of these disciplines and arguably no one individual can keep abreast of all fields. Consequently, we need to work together to solve problems, understanding our respective strengths, approaches and points of view.
This talk is about the cross-disciplinary work we have been doing in the Biomedical Data Science Laboratory in the UTS Australian Artificial Intelligence Institute, which is currently ranked 10th worldwide and first in Australia (AI Research Index, Software Policy and Research Institute). The BDS Lab and the Tumour Bank at The Children’s Hospital at Westmead have been working together since 2003, initially on two colour gene expression microarrays for acute lymphoblastic leukaemia, but latterly on stained tissue microarrays.
We will discuss some of the research we have done using image processing and neural networks on the data from stained tissue microarrays in paediatric neuroblastoma and rhabdomyosarcoma domains. Next we will talk about some of the outcomes of a novel co-learning workshop we ran to understand the drivers and inhibitors to cross-disciplinary research. Finally, we will show some of the recent work we have done in visualising spaces of data derived from patients, including a virtual reality system to explore oncology models.