As part of the Imaging ONEWORLD series, the focus of these lectures is on microscopy and image analysis methods and how to apply these to your research. Almost all aspects of imaging such as sample preparation, labelling strategies, experimental workflows, ‘how-to’ image and analyse, as well as facilitating collaborations and inspiring new scientific ideas will be covered. Speakers will be available for questions and answers. The organisers, core facility staff from the University of Cambridge, Gurdon Institute, MRC-LMB and the ICR/Royal Marsden Trust are also able to continue the discussion and provide advice on your imaging projects.
Scientific Organisers
Stefanie Reichelt
Stefanie Reichelt, PhD has been head of the light microscopy facility at the CRUK Cambridge Institute. The core provides state-of-the-art imaging resources, training courses for scientists and students and develop new imaging systems as well as user-friendly analysis and acquisition tools for specific research applications. Stefanie is now Public Engagement Manager for the Biomedical Schools and teaches academically at Cambridge University, in scientific workshops and out-reach events. (http://cargocollective.com/StefanieReichelt)
- Email:
- sr411@cam.ac.uk
Alessandro Esposito
Dr Alessandro Esposito obtained a PhD in Biophysics in 2006 working at the University of Utrecht and the European Neuroscience Institute in Goettingen for which he was awarded the ‘Sergio Ciani’ award by the Italian Society of Pure and Applied Biophysics. At the University of Cambridge, he then developed novel analytical tools contributing to redefining models of red blood cells homeostasis infected by P. falciparum (malaria). In recognition of his early work, in 2009 Alessandro was awarded a Life Science Interface fellowship by the EPSRC to establish foster the development of heavily multiplexed biochemical imaging. Soon after he moved to the MRC Cancer Unit where he lead the ‘Systems Microscopy initiative’ and retrained in cancer biology. During these years, Alessandro’s work developed into two research streams: i) the study of cellular responses to DNA damage and mutations in signalling pathways and ii) the innovation of biochemical imaging technologies. His team contributed to revealing the vast cell-to-cell variability in stress responses of genetically identical cells, a feature of biological systems that hinder the efficacy of disease management and therapeutic efficacy. Since 2019, Alessandro leads a transdisciplinary research programme at the MRC Cancer Unit in Cambridge devoted to understanding how DNA damage and mutations in KRAS derange homeostatic programmes leading to cancer. His group combines multi-omics data with single-cell biochemical imaging techniques aiming to achieve a deeper understanding of cancer phenotypes during the earliest stages of carcinogenesis, with particular attention to cell-to-cell variability of non-genetic origin and cell-to-cell communication.
Kirti Prakash
Kirti Prakash is a computer scientist by training (Bachelors and Masters degree) but a biologist at heart (PhD degree). Kirti aspires to be an inventor and develop new imaging tools for cell biology and neuroscience. Kirti did his Masters in Computer Science from Aalto University (Finland) and PhD in Biology from Heidelberg University (Germany). During his PhD, he developed a new method to image DNA which led to the first high-resolution images of the epigenetic landscape of meiotic chromosomes and mechanisms behind chromosome condensation. The doctoral research earned him several awards including Springer Best PhD Thesis Prize. After his PhD, he did a couple of postdocs at Carnegie Institution for Science (USA) and University of Cambridge (UK). The primary highlights of his research here were laser-free superresolution microscopy and development of a high-content imaging pipeline to quantify single-cell gene expression. Formerly at the National Physical Laboratory (NPL), and currently working at the Institute for Cancer Research (ICR) and Royal Marsden Trust, he is working on microscope development and image analysis.
Nick Barry
An optical physicist and specialist in light microscopy and head of the Light Microscopy facility at the MRC Laboratory of Molecular Biology, University of Cambridge.
BaSiCPy: an accelerated python program for uneven illumination correction of microscopy images
Microscopes are fundamental tools in life science. Despite their abundance and importance in research, accurate and efficient quantification of microscopy data is far from being straightforward. In microscopy images, real biologically-related signals could be mingled with experimental noise and batch variations, so methods need to be developed to disentangle signals from noise. Moreover, unlike natural images in the computer vision field, where annotations are readily available, ground-truth labels in microscopy images are usually more challenging to obtain as they often require expert annotation. In many scenarios, there is no ground-truth ‘perfect’ image at all.
In this talk, we will present a few examples of our developed AI methods to address these unique challenges in microscopy data, covering image preprocessing and enhancement, efficient learning through self-supervision and semi-supervision, integration of physical-based models, and convolutional neural networks. Particularly, we will focus on BaSiCPy, a Python package to correct uneven illumination for microscopy images, as one exemplary case of no ground-truth images with perfect illumination. The BaSiCPy is our recent Python implementation of the BaSiC algorithm, which estimates the multiplicative and additive components of the uneven illumination effects by L1 regularization. Compared to the original BaSiC which could only run on CPUs, our current BaSiCPy builds upon the JAX library supporting GPUs and TPUs, giving it a speed boost and making it more suited for large-scale image processing. In addition to implementation optimization, we also updated the algorithm itself, including a few recently developed extensions in BaSiCPy, such as employing rough foreground/background segmentation masks to improve illumination correction on images with correlated foreground, and shadowing correction of 3D image volumes. We also built a plugin for the napari image viewer, which makes it possible to perform the analysis interactively without coding.
Dr Tingying Peng
