Integrated Cancer Medicine Seminar Series #2 AI in Oncology
Event Information
About this event
This series comprises speakers who work within the area of integrated cancer medicine both in the research space and in the clinical space. Talks will be 15-30 minutes in length with time for Q&A afterwards on zoom. We want to target a broad audience who are interested in how integrating patient data can lead to revolutionising cancer care by re-inventing their treatment pathway.
ICM Seminar Series #2 - AI in Oncology
In the second seminar of the series hosted by Dr Leonardo Rundo from the CRUK Cambridge Centre, Dr Alex Graudenzi from the Institute of Bioimaging and Molecular Physiology, Consiglio Nazionale delle Ricerche (IBFM-CNR), Milan will talk about Reconstruction of cancer evolution models from longitudinal single-cell sequencing data and Dr Nikola Simidjievski from the Mark Foundation Institute for Integrated Cancer Medicine, University of Cambridge will talk about Machine learning for integrative and interpretable cancer data-analysis .
Reconstruction of cancer evolution models from longitudinal single-cell sequencing data
The rise of longitudinal single-cell sequencing experiments on patient-derived cell cultures, xenografts and organoids is opening exciting new opportunities to track cancer evolution, assess the efficacy of therapies and identify resistant subclones.
In this seminar, I will introduce LACE, the first algorithmic framework that processes single-cell mutational profiles from samples collected at different time points to reconstruct longitudinal models of cancer evolution.
On simulations, LACE outperforms state-of-the-art methods for both bulk and single-cell sequencing data. Moreover, as the results are robust with respect to data-specific errors, LACE is effective also with mutational profiles generated by calling variants from (full-length) scRNA-seq data, and this allows one to investigate the relation between genomic and phenotypic evolution of tumors at the single-cell level.
As a proof of principle, I will finally show the application of LACE to a longitudinal scRNA-seq dataset of patient-derived xenografts of BRAF-mutant melanomas, dissecting the impact of BRAF/MEK-inhibition on clonal evolution, also in terms of clone-specific gene expression dynamics.
Alex Graudenzi is a Tenured Researcher of the Institute of Bioimaging and Molecular Physiology, Consiglio Nazionale delle Ricerche (IBFM-CNR), Milan.
He received his Ph.D. in computational modeling and simulation in 2010. Since then, he has been Research Associate, Assistant Professor and Visiting Scientist at several important Research Centers and Universities.
He is author of 60+ publications on indexed international journals and conference proceedings, recipient of several research awards, scientific collaborator in many international projects, co-developer of 15+ tools for computational biology, and organizer, chair and program committee member of numerous international conferences.
He works at the boundaries of Bioinformatics, Complex Systems and Artificial Intelligence to investigate the properties of biological systems and, especially, of cancer evolution.
Machine learning for integrative and interpretable cancer data-analysis
Cancer research produces large amounts of complex heterogeneous data on different scales (multi-omic, clinical, image etc.). These data, besides originating from different sources, typically are high-dimensional (and limited to few observations), noisy, and scarce. In response, machine-learning approaches that leverage such integrative data analysis scenarios can provide better understanding of the underlying mechanisms of a biological process and ultimately lead to more accurate cancer diagnosis, prognosis and treatment planning.
The challenges of many current machine learning approaches for integrative analysis are their complexity and lack of interpretability. In this talk we are going to present recent approaches that are able to address these challenges from different perspectives. First we will discuss methods for efficient integration of heterogeneous data. Second, we will also focus on methods that also take into consideration domain expert knowledge of biologically relevant information. Finally we will present current work on integrative and explainable pipelines for cohort- and patient-wise data analyses that facilitate the bench-to-bedside process in clinical decision-support tools for cancer medicine.
Nikola Simidjievski is a Senior Research Associate at the Department of Computer Science and Technology, working with Prof. Mateja Jamnik and Prof. Pietro Liò. His research interests are at the intersection of machine learning and natural sciences (medicine, biology and neuroscience). More specifically, he is interested in various topics of Bayesian inference for development of novel machine learning algorithms for integrative data analysis and their application in cancer medicine. Simidjievski has experience in computational scientific discovery, and is quite keen on machine learning for modelling dynamic systems.