Convolutional Neural Networks for Classifying Chromatin Morphology in Live-Cell Imaging

Abstract

Chromatin is highly structured, and changes in its organization are essential in many cellular processes, including cell division. Recently, advances in machine learning have enabled researchers to automatically classify chromatin morphology in fluorescence microscopy images. In this protocol, we develop user-friendly tools to perform this task. We provide an open-source annotation tool, and a cloud-based computational framework to train and utilize a convolutional neural network to automatically classify chromatin morphology. Using cloud compute enables users without significant resources or computational experience to use a machine learning approach to analyze their own microscopy data.

Keywords: Cell cycle; Computational biology; Image analysis; Live-cell imaging; Machine learning.