A matlab toolbox for ultrastructural analysis of filament networks in cryo-ET data

Filamentous assemblies are ubiquitous in cells and tissues and play a key role in many biological phenomena such as cell integrity, cell motility, tissue formation, and maintenance. Manipulation of the associated regulatory proteins in these structures results in changes in their ultrastructural properties. Thus, characterization of these assemblies in wild-type or various experimental conditions can provide an insight into the role of different players. One major challenge is that sufficiently large datasets are usually required to demonstrate subtle ultrastructural differences between features of multiple experimental groups. Recent advances in high-throughput cryo-electron tomography (cryo-ET) methods facilitate the acquisition of such large datasets. These datasets should be transformed into relevant biological information by deriving parameters for filament networks through ultrastructural analysis. We have developed a MATLAB-based computational toolbox that enables the semi-automatic quantification of filament networks from large cryo-ET datasets. This toolbox allows cross-comparison of experimental conditions by deriving a large number of ultrastructural parameters describing filamentous assemblies. The GUI of the toolbox also offers various options for the curation and analysis of data. The functionality of the workflow is tested on manipulated lamellipodial actin networks and parallel actin filament arrays in filopodia or microspikes. It can identify key structural differences between experimental conditions and facilitate the in-depth analysis of large ultrastructural datasets. The computational toolbox introduced here can in principle be used for the analysis of any filamentous network in ET data, such as the described branched networks within lamellipodia or quasi-parallel structures in filopodia or microspikes.