Start date: Week 39 (24th September 2024)
Speaker: Daniel Fürth (SciLifeLab Fellow; Tuesday 24th September 8:00-10:00 CEST)
This module will explore advanced tools for single-molecule imaging and programmable bioconjugation to study cell communication. It will cover essential computational methods for image analysis of single molecules in both fixed and living cells. Participants will learn about the segmentation of individual molecules in 2D and 3D using U-Net and ResNet deep learning architectures, and how to track these molecules over time using DeepSORT, an enhanced version of the SORT algorithm.
The session will also discuss transforming images of tissue sections onto a known anatomical reference atlas using rigid and non-rigid transformations. Participants will explore point-cloud registration algorithms like Thin-Plate Splint (TPS) registration and Coherent-Point Drift (CPD), comparing their relative merits. Additionally, the lecture will introduce valuable resources, database structures, and ontologies relevant to creating anatomical atlases.
Computer Lab: The lab session will involve hands-on exercises using Jupyter notebooks to apply the discussed computational tools for analyzing fluorescent microscopy data. Participants will practice segmenting and tracking single molecules, as well as performing tissue atlas registration.