For almost twenty years the Danuser lab has developed a wide range of software for live microscopy image analysis, including particle tracking. In a landmark paper in 2008 the lab introduced a versatile algorithmic platform for the robust tracking of dense fields of interacting and partially unstable particle images. The software, called u-track, has ever since been defining the state-of-art for particle tracking, which is an ubiquitous task in the analysis of live microscopy data, from super-resolution and single molecule studies to studies of tissue development and homeostasis. The algorithmic core of u-track has been adopted by the widely-used TrackMate/ImageJ and CellProfiler open science software packages, as well as by commercial software platforms like Imaris and Amira. Nonetheless, u-track retained its popularity as standalone, specialized software package for particle tracking applications because of numerous advanced features these adoptions did not include. The lab has made a significant effort in maintaining u-track as a user-friendly package and in supporting the user community through consultations. Over the past four years the lab has expanded u-track's functionality for the analysis of time-lapse 3D image stacks, which are becoming more and more ubiquitous thanks to the rapid development of light-sheet microscopy. Especially with u-track3D many labs will lack the necessary computing infrastructure for the analysis of these massive data sets, and installation of a specialized open source software on a departmental or institutional high-performance computing infrastructure can be cumbersome. Moreover, the burden on our lab of maintaining the u-track platform with appropriate backward and forward version compatibility has become significant. Therefore, we propose a one year project supplementing the recently funded MIRA grant R35 GM136428-01 to migrate the current u-track and u-track3D software packages into a containerized pipeline, called u-trackAll, for platform-agnostic cloud-based and/or local computation. Specifically we will, Aim 1: Develop u-trackAll combining u-track and u-track3D into a generic particle tracking platform;
Aim 2 : Develop u-trackAll as a container-mediated, platform-agonistic workflow;
Aim 3 : Develop strategies for hybrid (local + cloud) interactive data visualization. The work will be performed by a project team composed of the current Danuser lab software engineer and a software engineer in UT Southwestern's Bioinformatics Core Facility, which employs a group of software developers dedicated to the assembly of open science software workflows.
We propose to turn our widely accessed u-track software into a robust and easily maintained open science platform for image particle tracking, which is an ubiquitous task in the analysis of live microscopy data, from super-resolution and single molecule studies to studies of tissue development and homeostasis. The platform will support local and cloud-based computing alike.
