CytoData Symposium 2022

Featured Speakers

CytoData Symposium 2022: Featured Speakers


Emma Lundberg, PhD | Stanford University – Keynote Speaker 

Talk title: Mapping the spatiotemporal proteome architecture of human cells 
Abstract: Biological systems are functionally defined by the nature, amount and spatial location of the totality of their proteins. We have generated an image-based map of the subcellular distribution of the human proteome and showed that there is great complexity to the subcellular organization of the cell. As much as half of all proteins localize to multiple compartments, giving rise to potential pleiotropic effects, and around 20% of the human proteome shows spatiotemporal variability. Our temporal mapping results shows that cell cycle progression explains less than half of all temporal protein variability, and that most cycling proteins are regulated post-translationally, rather than by transcriptomic cycling. This work is critically dependent on computational image analysis, and I will discuss machine learning approaches for classification of spatial subcellular patterns and how such embeddings can be used to build multi-scale models of cell architecture. In summary, I will demonstrate the importance of spatial proteomics data for improved single cell biology and present how the freely available Human Protein Atlas database (www.proteinatlas.org) can be used as a resource for life science.

Bio: Dr. Lundberg is Associate Professor of Bioengineering and Pathology at Stanford University, and Director of the Cell Atlas, of the Human Protein Atlas program. In the interface between bioimaging, proteomics and artificial intelligence her research aims to define the spatiotemporal subcellular architecture of the human proteome, to understand how variations in protein expression patterns contribute to cellular function and disease. Dr. Lundberg has a keen interest in citizen science using computer games, and has engaged over 300,000 gamers to help her classify images in the first ever MMO citizen science computer game. 


Pedro Falé Ferreira, MSc | ETH Zürich 

Talk title: Methods to understand tumor heterogeneity from single-cell data
Abstract: Cancer arises and evolves by the accumulation of somatic mutations that provide a selective advantage. The interplay of mutations and their functional consequences shape the evolutionary dynamics of tumors and contribute to different clinical outcomes. In this talk I discuss computational methods to analyse transcriptomes of cancer cells and their relationship to the evolution of somatic copy number aberrations in the tumor. I present novel statistical models to identify hierarchical gene expression signatures from single-cell RNA-sequencing data, and to integrate single-cell transcriptomes with copy number evolutionary trees. These methods provide rich descriptions of tumor heterogeneity and complement genotype-only analyses of tumor evolution. 

Bio: I am a doctoral student at the Computational Biology Group at the department of Biosystems Science and Engineering of ETH Zürich in Basel, Switzerland. I am generally interested in analyzing large data sets using statistical models. More specifically, in my research I develop high-dimensional generative models and learning algorithms to extract information from large- scale single- cell sequencing data and reconstruct the evolution of tumors. I am supervised by Niko Beerenwinkel and Jack Kuipers. I obtained a master’s degree in electrical engineering at the Instituto Superior Técnico in Lisbon, Portugal, with a major in communication systems.


Uri Manor, PhD | Salk Institute for Biological Studies

Talk title: Deep Learning-Based Tools for Tracking Organelle Mobility in Live Neurons 
Abstract: Unfortunately, many fundamental aspects of neuronal cell biology remain poorly understood. For example, while many neurodegenerative diseases are associated with genes involved in organelle dynamics, our understanding of neuronal organelle behavior and function in health and disease is incomplete, largely due to the lack of tools to measure and interpret changes to organelle dynamics over time in living neurons, where organelles must regularly undergo transport over relatively long distances. Measuring organelle dynamics requires advanced imaging equipment with high spatiotemporal resolution, and similarly cutting-edge image analysis tools with sufficient precision and accuracy for these tasks. Towards this end, we are developing deep learning-based image analysis tools that enable us to accurately track the movement of mitochondria and lysosomes. We have applied these tools towards neurons derived from patients with Charcot-Marie-Tooth disease, revealing a significant defect in organelle mobility, which we hypothesize to be a key underlying cause of the disease. 

Bio: The Manor Lab develops new methods and tools for studying cellular dynamics with nanometer precision (a sheet of paper is about 100,000 nanometers thick). This includes artificial-intelligence-based computational approaches (deep learning) that integrate data from microscopes to increase image resolution, sensitivity, and collection speed beyond what’s possible with any other existing method. The Manor Lab also develops genetic and molecular tools that facilitate the monitoring and manipulation of cellular structures implicated in diseases including neurodegenerative diseases and hearing loss. Using these advanced technologies, the Manor Lab connects anatomy to function. Their research advances scientists’ understanding of these cellular processes and ultimately helps discover and create new therapies for treating these conditions. 


Lucas Pelkmans, PhD | University of Zurich

Talk title: "Multimodal perception links cellular state to decision making in single cells"
Abstract: Individual cells make decisions that are adapted to their internal state and surroundings, but how cells can reliably do this remains unclear. To study the information processing capacity of human cells, we conducted multiplexed quantification of signaling responses and markers of the cellular state. Signaling nodes in a network displayed adaptive information processing, which led to heterogeneous growth factor responses and enabled nodes to capture partially nonredundant information about the cellular state. Collectively, as a multimodal percept this gives individual cells a large information processing capacity to accurately place growth factor concentration within the context of their cellular state and make cellular state-dependent decisions. Heterogeneity and complexity in signaling networks may have coevolved to enable specific and context-aware cellular decision-making in a multicellular setting.

Bio: Lucas Pelkmans did his PhD in Biochemistry at the ETH Zurich. He was then a postdoctoral fellow at the Max Planck Institute of Molecular Cell Biology and Genetics in Dresden, Germany. At the age of 29, he became assistant professor at the Institute of Molecular Systems Biology of the ETH Zurich and was elected the Ernst Hadorn-endowed Chair at the University of Zurich 5 years later in the Department of Molecular Life Sciences, where he is currently a full professor. Lucas Pelkmans was chairman of the Scientific Executive Board of SystemsX.ch from 2013-2018 and was elected member of the European Molecular Biology Organisation (EMBO) in 2015. He is an inventor on several patents in image-based systems biology and is a co-founder of 3VBiosciences (now Sagimet Biosciences) and Apricot Therapeutics. He obtained the ETH medal, won the European Young Investigator Award, and received ERC junior, consolidator, and advanced grants.


Susanne Rafelski, PhD | Allen Institute for Cell Science

Talk title: Integrated intracellular organization and its variations in human iPS cells 
Abstract: Understanding how a subset of expressed genes dictates cellular phenotype is an enormous challenge due to the large numbers of molecules, their combinatorics, and the plethora of cellular behaviors they determine. We reduced this complexity by focusing on cellular organization, a key readout and driver of cell behavior, at the level of major cellular structures representing distinct organelles and functional machines, and generated the “hiPSC Single-Cell Image Dataset” with over 200,000 live cells in 3D spanning 25 major cellular structures. The scale and quality of this dataset permitted the creation of a generalizable analysis framework to convert raw image data of cells and their structures into dimensionally reduced, human-interpretable quantitative measurements and to facilitate data exploration. This framework embraces the vast cell-to-cell variability observed within a normal population, facilitates integration of cell-by-cell structural data, and permits quantitative analyses of distinct, separable aspects of organization within and across different cell populations. We found that the integrated intracellular organization of interphase cells was robust to the wide range of cell shape variations in the population, that the average locations of some structures became polarized in cells at the edges of colonies while maintaining the “wiring” of their interactions with other structures, and that, in contrast, structure location changes during early mitotic reorganization were accompanied by changes in their wiring.

Bio: Susanne obtained her B.S. in Biochemistry and Molecular & Cellular Biology with an additional emphasis in Mathematics from the University of Arizona. Susanne then completed her Ph.D. in Biochemistry at Stanford University, followed by a postdoc at the Center for Cell Dynamics at the Friday Harbor Labs, University of Washington, where she learned computational modeling approaches. Susanne then initiated her current research program on mitochondrial structure-function as a postdoc at UCSF, where she developed 3D microscopy and image analysis methods to quantify mitochondrial morphology and applied these to investigate mitochondrial size control regulation. As a model system for intracellular organization, the Rafelski lab extended this work to studying the size, topology, and function of mitochondrial networks in budding yeast and mammalian cells. Prior to joining the Allen Institute, Susanne was an Assistant Professor in the Department of Developmental and Cell Biology, the Department of Biomedical Engineering, and the Center for Complex Biological Systems at UC Irvine.  

Susanne started at the Institute in 2016 as the Director of Assay Development where she contributed to developing and implementing the overall scientific direction of the Institute. In January 2020 she was promoted to Deputy Director at the Allen Institute for Cell Science where she continues to guide their scientific and strategic goals. 


Ana Sanchez-Fernandez, MSc | Johannes Kepler University

Talk title: Cross-modal contrastive learning of microscopy image- and structure-based representations of molecules
Abstract: Contrastive learning for self-supervised representation learning has brought a strong improvement to many application areas. With the availability of large collections of unlabeled data, contrastive learning of language and image representations has shown impressive results. The methods CLIP and CLOOB have demonstrated that the learned representations are highly transferable to a diverse set of tasks when trained on data from two different domains. In drug discovery, similar large, multi-modal datasets comprising both cell-based microscopy images and chemical structures of molecules are available. However, contrastive learning has not been used for this type of data, although transferable representations could be a remedy for the time-consuming and cost-expensive label acquisition in this domain. We present CLOOME, a contrastive learning method for image- and structure-based representations of small molecules for drug discovery. On the benchmark dataset Cell Painting, we demonstrate its ability to learn proficient representations by performing linear probing for activity prediction tasks.

Bio: Ana Sanchez-Fernandez is a PhD student at Johannes Kepler University Linz (Austria). She is a fellow within the Marie Curie European project Advanced Machine Learning for Innovative Drug Discovery (AIDD). Her project focuses on representation learning for molecules making use of their chemical structures and corresponding microscopy imaging data. She earned a BSc in Biochemistry from the University of Murcia and a MSc in Bioinformatics from Pompeu Fabra University.


David Van Valen, MD, PhD | California Institute of Technology

Talk title: Everything as code
Abstract: Biological systems are difficult to study because they consist of tens of thousands of parts, vary in space and time, and their fundamental unit—the cell—displays remarkable variation in its behavior. These challenges have spurred the development of genomics and imaging technologies over the past 30 years that have revolutionized our ability to capture information about biological systems in the form of images. Excitingly, these advances are poised to place the microscope back at the center of the modern biologist’s toolkit. Because we can now access temporal, spatial, and “parts list” variation via imaging, images have the potential to be a standard data type for biology. 

For this vision to become reality, biology needs a new data infrastructure. Imaging methods are of little use if it is too difficult to convert the resulting data into quantitative, interpretable information. New deep learning methods are proving to be essential to reliable interpretation of imaging data. These methods differ from conventional algorithms in that they learn how to perform tasks from labeled data; they have demonstrated immense promise, but they are challenging to use in practice. The expansive training data required to power them are sorely lacking, as are easy-to-use software tools for creating and deploying new models. Solving these challenges through open software is a key goal of the Van Valen lab. In this talk, I describe DeepCell, a collection of software tools that meet the data, model, and deployment challenges associated with deep learning. These include tools for distributed labeling of biological imaging data, a collection of modern deep learning architectures tailored for biological image analysis tasks, and cloud-native software for making deep learning methods accessible to the broader life science community. I discuss how we have used DeepCell to label large-scale imaging datasets to power deep learning methods that achieve human level performance and enable new experimental designs for imaging-based experiments. 

Bio: David Van Valen PhD is faculty in the Division of Biology and Bioengineering at Caltech. His research group's long-term interest is to develop a quantitative understanding of how living systems process, store, and transfer information, and to unravel how this information processing is perturbed in human disease states. To that end, his group leverages and pioneers the latest advances in imaging, genomics, and machine learning to produce quantitative measurements with single-cell resolution as well as predictive models of living systems. Prior to joining the faculty, he studied mathematics (BS 2003) and physics (BS 2003) at the Massachusetts Institute of Technology, applied physics (PhD 2011) at Caltech, and medicine at the David Geffen School of Medicine at UCLA (MD 2013)


Assaf Zaritsky, MSc, PhD | Ben-Gurion University of the Negev

Talk title: Visual interpretability of deep learning models in cell imaging 
Abstract: Deep learning has emerged as a powerful technique to identify hidden patterns in complex cell imaging data, but is criticized for lacking the ability to offer meaningful explanations of which image properties drive the network prediction. I will present methods to "reversed engineered" neural networks by generating synthetic images in a controlled manner to amplify subtle image properties that critically define the model’s classification and demonstrate applications in predicting melanoma metastatic aggressiveness, in vitro fertilization success, and image-image translation. 

Bio: Assaf Zaritsky is a Senior Lecturer in the Department of Software and Systems Information Engineering at Ben-Gurion University of the Negev (BGU), Israel. He received his B.Sc. and M.Sc. degrees in computer science at BGU, and Ph.D. in computer science at Tel Aviv University. He was a postdoctoral fellow at UT Southwestern Medical Center and joined BGU in October 2018. His research is in computational cell dynamics, at the interface of data science and cell biology with specific focus on multicellular information processing and interpretable AI in cell imaging. Lab website: https://www.assafzaritsky.com/, Twitter: https://twitter.com/AssafZaritsky