DeftPunk: an open-source program for detecting topological defects and measuring nematic splay-bend anisotropy.

Abstract

Here, we present an open Python-based program with a graphical interface to characterise topological defects from two-dimensional time-lapses. This program computes the nematic orientation field, identifies defect positions, topological charges, and defect orientations, and estimates the splay-bend anisotropy parameter. The workflow is controlled by three detection parameters adjustable through the interface: the feature size, the order threshold and the phase contour radius. Validation on synthetic datasets with known ground truth yields less than 0.2% false positives and negatives and positional errors comparable to the resolution scale, and the absolute error of the splay-bend anisotropy is 0.03. In addition, we explain a correlation between the detected and input splay-bend anisotropy and provide a calibration curve. Next, we demonstrate applicability across biological systems spanning multiple scales. DeftPunk provides a versatile tool to characterise topological defects and infer elastic anisotropies from microscopy images, opening the way to systematic and quantitative investigations of active nematic dynamics in living systems.