Design, Analysis, and Translation of 3D Hydrodynamic Tweezer Microeddies

Abstract

The drive to understand the behavior and dynamics of single cells and their connection to population-based properties has prompted the microfluidic community to develop diverse micro-trapping array systems. We present work on the use of microeddy arrays for trapping, counting, and characterizing microscale objects (particles and cells). We explore the effects of nine distinct device geometries on microeddy flow traits, and show how the eddy number, shape, symmetry, and strength are controlled by the device geometry. Microparticle trapping stability and trapping site depends on the frequency of flow oscillations as well as the device geometry. When placed in arrays, microeddies provide a high throughput platform for microparticle and cell trapping. Proper statistical design is critical for quantitatively linking single-cell measurements to population behavior. We describe a general procedure for evaluating data quality, detection, and determination limits when using arrayed trapping devices that load a homogenized mixture of micro-objects, and then trap those objects. A series of successively stringent statistical tests are used to evaluate the operational domain where micro-object trapping is described by a Poisson distribution. When particle titer (particles/ml) is well above the trap titer (traps/ml), finite size effects and particle masking cause deviations from Poisson behavior. We use the statistical properties of the array to determine the detection limit for rare objects and uncertainty in quantitative measurements. In the dilute trapping limit, we also show that counting the empty traps is an effective method for determining titer. Statistical design is a powerful tool for creating combinations of different micro-objects, for example to study cell-cell interactions. With known trapping statistics, we show that the assembly of a predictable distribution of singlet, doublet, triplet and higher order particle clusters is possible, along with predictable sub-population ratios in each doublet or larger cluster. Because each microeddy is hydrodynamically-isolated from the others, this platform seems well suited for studying cell-cell interactions via paracrine signaling. The statistical tools and experimental design we present here can be applied to different kinds of microtrapping arrays, and should broadly help guide device system designers to support biological research.