SOAX: A software for quantification of 3D biopolymer networks

Ting Xu, Dimitrios Vavylonis, Feng-Ching Tsai, Gijsje H. Koenderink, Wei Nie, Eddy Yusuf, I-Ju Lee, Jian-Qiu Wu, Xiaolei Huang
2015 Scientific Reports  
Filamentous biopolymer networks in cells and tissues are routinely imaged by confocal microscopy. Image analysis methods enable quantitative study of the properties of these curvilinear networks. However, software tools to quantify the geometry and topology of these often dense 3D networks and to localize network junctions are scarce. To fill this gap, we developed a new software tool called "SOAX", which can accurately extract the centerlines of 3D biopolymer networks and identify network
more » ... ions using Stretching Open Active Contours (SOACs). It provides an open-source, user-friendly platform for network centerline extraction, 2D/3D visualization, manual editing and quantitative analysis. We propose a method to quantify the performance of SOAX, which helps determine the optimal extraction parameter values. We quantify several different types of biopolymer networks to demonstrate SOAX's potential to help answer key questions in cell biology and biophysics from a quantitative viewpoint. N etwork structures made of filamentous biopolymers are ubiquitous among biological systems. Biophysicists and cell biologists routinely use static and time-lapse confocal fluorescence microscopy to image intracellular networks of actin filaments 1,2 and microtubules 3,4 as well as extracellular polymers such as fibrin 5,6 , both in vitro and in live cells. To gain insight in the structural, dynamical, and mechanical properties of these networks and to understand the mechanisms of their formation requires image analysis methods for automated quantification of massive image datasets. However, user-friendly, flexible, and transparent 7 software tools to reliably quantify the geometry and topology of these (often dense) networks and to localize network junctions in 3D are scarce. Previous methods for extracting biopolymer network structures include morphological thinning of a binary segmentation 8-11 or a computed tubularity map 12,13 , Radon transform 14 and template matching 15,16 . However, most of these methods extract disconnected points (i.e. pixels) on centerlines without inferring network topology and they have not been implemented as part of a software platform. One available software tool is "Network Extractor" (, which finds one-pixel wide 3D network centerlines by thresholding and thinning a tubularity map. Thresholding results, however, can suffer from inhomogeneous signal-to-noise ratio (SNR). Other software for extracting curvilinear network structure are designed for neuronal structures 17-20 . Vaa3D-Neuron 19 ( is a semi-automatic neuron reconstruction and quantification tool which requires the user to pinpoint the end points of a neuronal tree so that a minimal path algorithm can reconstruct the structure. The Farsight Toolkit ( also contains 3D neuron tracing and reconstruction software command-line modules 21, 22 . To fill this gap in available software, here we provide an open source program, SOAX, designed to extract the centerlines and junctions of biopolymer networks such as those of actin filaments, microtubules, and fibrin, in the presence of image noise and unrelated structures such as those that appear in images of live cells. SOAX provides quantification and visualization functions in an easy-to-use user interface. The underlying method of SOAX is the multiple Stretching Open Active Contours (SOACs) method that was proposed to extract the 3D meshwork of actin filaments imaged by confocal microscopy 23 . Here we implement this method in SOAX and apply it generally to different types of biopolymer networks. While the SOAX method is robust against noise, its parameters need to be adjusted depending on the type of biopolymer and the image SNR. Parameters for actin filaments were previously chosen empirically 23 . Here we provide a new method to evaluate the accuracy of the network extraction results and find a small set of candidate optimal solutions for the user to OPEN SUBJECT AREAS: FLUORESCENCE IMAGING IMAGE PROCESSING CELLULAR IMAGING CYTOSKELETON
doi:10.1038/srep09081 pmid:25765313 pmcid:PMC4357869 fatcat:zssc7owqlbfsfo3hqntn2jghoa