High Throughput Screening towards the Rational Design of Active Matter

We introduce sample-agnostic techniques to screen microscopy data from active matter systems for three properties of interest: ability to transmit force, significant contractile motion, and lifetime of structures. Screening of large data sets can be laborious, and may slow optimization of experimental protocols, and so means of rapidly identifying promising specimens are highly desirable. We have developed computational tools to apply efficient heuristics to series of images to identify samples which may have a desired trait. Because all techniques are designed to be sample agnostic, all that is required is a set of image intensity data at regular time intervals. We thus anticipate our software will be of interest to the broader soft matter community. We assess the possibility of force transmission by testing for correlation of velocity directions over large distances, a necessary, though not sufficient condition. We have found a decreasing correlator corresponds with the breakup of anactomyosin network into small clusters. To check for contractility, we look for a decreasing characteristic decay length in image intensity autocorrelations. Finally, we quantify the persistence of the spatial structure of a sample by computing the largest void size in the intensity field.