Mammalian cell culture in nanoliter droplets of extracellular matrix has recently attracted interest as a platform for high throughput drug screening on 3D tissue models. Microfluidic droplet generation using basement membrane extract, such as Matrigel or Cultrex, is complicated by the fact that the viscosity is orders of magnitude higher than that of water. Consequently, conditions that produce quality water droplet generation can fail with basement membrane extract due to the higher capillary number. Here, a parametric study using a T-junction device demonstrates that higher viscosity can be abrogated by lower interfacial tension, which is modulated by surfactant concentration.

PIEZO1 is critical to numerous physiological processes, transducing diverse mechanical stimuli into electrical and chemical signals. Recent studies underscore the importance of visualizing endogenous PIEZO1 activity and localization to understand its functional roles. To enable physiologically and clinically relevant studies on human PIEZO1, we genetically engineered human induced pluripotent stem cells (hiPSCs) to express a HaloTag fused to endogenous PIEZO1. Combined with advanced imaging, our chemogenetic platform allows precise visualization of PIEZO1 localization dynamics in various cell types. Furthermore, the PIEZO1-HaloTag hiPSC technology facilitates the non-invasive monitoring of channel activity across diverse cell types using Ca2+-sensitive HaloTag ligands, achieving temporal resolution approaching that of patch clamp electrophysiology. Finally, we use lightsheet microscopy on hiPSC-derived neural organoids to achieve molecular scale imaging of PIEZO1 in three-dimensional tissue. Our advances establish a platform for studying PIEZO1 mechanotransduction in human systems, with potential for elucidating disease mechanisms and targeted drug screening.

Systems biology models are useful models of complex biological systems that may require a large amount of experimental data to fit each model's parameters or to approximate a likelihood function. These models range from a few to thousands of parameters depending on the complexity of the biological system modeled, potentially making the task of fitting parameters to the model difficult - especially when new experimental data cannot be gathered. We demonstrate a method that uses structural biology predictions to augment systems biology models to improve systems biology models' predictions without having to gather more experimental data. Additionally, we show how systems biology models' predictions can help evaluate novel structural biology hypotheses, which may also be expensive or infeasible to validate.

Recent advances in generative deep learning have transformed small molecule design, but most methods lack biological systems context, focusing narrowly on specific protein pockets. We introduce a non-differentiable diffusion guidance method that integrates systems biology models, enhancing small molecule generation with pathway context. Using the Bone Morphogenetic Protein (BMP) pathway, we generate small molecules specific to one protein over competing proteins. This method enhances the precision and efficiency of small molecule drug discovery by incorporating systems biology insights into generative models.

PIEZO1, a mechanically-activated, Ca2+-permeable ion channel, regulates numerous biological processes. To understand its many physiological roles, it is vital to be able to visualize the channel and monitor its activity in the native cellular environment. However, current approaches for imaging PIEZO1, such as tagging the channel with fluorescent proteins, are limited by dim fluorophores prone to photobleaching. Furthermore, tools to visualize PIEZO1 activity in intact cells, including genetically encoded calcium indicators and organic Ca2+ indicators often exhibit slow kinetics or lack specificity.

Systems biology relies on mathematical models that often involve complex and intractable likelihood functions, posing challenges for efficient inference and model selection. Generative models, such as normalizing flows, have shown remarkable ability in approximating complex distributions in various domains. However, their application in systems biology for approximating intractable likelihood functions remains unexplored. Here, we elucidate a framework for leveraging normalizing flows to approximate complex likelihood functions inherent to systems biology models. By using normalizing flows in the Simulation-based inference setting, we demonstrate a method that not only approximates a likelihood function but also allows for model inference in the model selection setting. We showcase the effectiveness of this approach on real-world systems biology problems, providing practical guidance for implementation and highlighting its advantages over traditional computational methods.

Simulation-based inference (SBI) methods tackle complex scientific models with challenging inverse problems. However, SBI models often face a significant hurdle due to their non-differentiable nature, which hampers the use of gradient-based optimization techniques. Bayesian Optimal Experimental Design (BOED) is a powerful approach that aims to make the most efficient use of experimental resources for improved inferences. While stochastic gradient BOED methods have shown promising results in high-dimensional design problems, they have mostly neglected the integration of BOED with SBI due to the difficult non-differentiable property of many SBI simulators. In this work, we establish a crucial connection between ratio-based SBI inference algorithms and stochastic gradient-based variational inference by leveraging mutual information bounds. This connection allows us to extend BOED to SBI applications, enabling the simultaneous optimization of experimental designs and amortized inference functions. We demonstrate our approach on a simple linear model and offer implementation details for practitioners.

Alternative computing approaches that interface readily with physical systems are well suited for embedded control of those systems. We demonstrate finite state machines implemented as pneumatic circuits of microfluidic valves and use these controllers to direct microfluidic liquid handling procedures on the same chip. These monolithic integrated systems require only power to be supplied externally, in the form of a vacuum source. User input can be provided directly to the chip by covering pneumatic ports with a finger. State machines with up to four bits of state memory are demonstrated, and next-state combinational logic can be fully reprogrammed by changing the hole-punch pattern on a membrane in the chip. These pneumatic computers demonstrate a framework for the embedded control of physical systems and open a path to stand-alone lab-on-a-chip devices capable of highly complex functionality.

Systems biology seeks to create math models of biological systems to reduce inherent biological complexity and provide predictions for applications such as therapeutic development. However, it remains a challenge to determine which math model is correct and how to arrive optimally at the answer. We present an algorithm for automated biological model selection using mathematical models of systems biology and likelihood free inference methods. Our algorithm shows improved performance in arriving at correct models without a priori information over conventional heuristics used in experimental biology and random search. This method shows promise to accelerate biological basic science and drug discovery.

Dry film photoresists are widely employed to fabricate high-aspect-ratio microstructures, such as molds for microfluidic devices. Unlike liquid resists, such as SU-8, dry films do not require a cleanroom facility, and it is straightforward to prepare uniform and reproducible films as thick as 500 µm. Multilayer patterning, however, can be problematic with dry film resists even though it is critical for a number of microfluidic devices. Layer-to-layer mask alignment typically requires the first layer to be fully developed, making the pattern visible, before applying and patterning the second layer. While a liquid resist can flow over the topography of previous layers, this is not the case with dry film lamination. We found that post-exposure baking of dry film photoresists can preserve a flat topography while revealing an image of the patterned features that is suitable for alignment to the next layer. We demonstrate the use of this technique with two different types of dry film resist to fabricate master molds for a hydrophoresis size-sorting device and a cell chemotaxis device.

Microfluidic droplet generation typically entails an initial stabilization period on the order of minutes, exhibiting higher variation in droplet volume until the system reaches monodisperse production. The material lost during this period can be problematic when preparing droplets from limited samples such as patient biopsies. Active droplet generation strategies such as antiphase peristaltic pumping effectively reduce stabilization time but have required off-chip control hardware that reduces system accessibility. We present a fully integrated device that employs on-chip pneumatic logic to control phase-optimized peristaltic pumping. Droplet generation stabilizes in about a second, with only one or two non-uniform droplets produced initially.

Microfluidic devices are traditionally monitored by bulky and expensive off-chip sensors. We have developed a soft piezoresistive sensor capable of measuring micron-level strains that can be easily integrated into devices via soft lithography. We apply this sensor to achieve fast and localized monitoring of pressure, flow, and valve actuation.

Biological signaling pathways based upon proteins binding to one another to relay a signal for genetic expression, such as the Bone Morphogenetic Protein (BMP) signaling pathway, can be modeled by mass action kinetics and conservation laws that result in non-closed form polynomial equations. Accurately determining parameters of biological pathways that represent physically relevant features, such as binding affinity of proteins and their associated uncertainty, presents a challenge for biological models lacking an explicit likelihood function. Additionally, parameterizing non-closed form biological models requires copious amounts of data from expensive perturbation-response experiments to fit model parameters. We present an algorithm (SBIDOEMAN) for determining optimal experiments and parameters of systems biology models with implicit likelihoods. We evaluate our algorithm using simulations of held-out true parameter values and demonstrate an improvement in the rate of accurate parameter inference over random and equidistant experimental designs when evaluated on two simple models of the BMP signaling pathway with an implicit likelihood function.

We present a simple method to pattern multiple cell populations inside a microfluidic channel. The microchannel is partially filled with a cell suspension, and the position of the liquid boundary remains pinned by surface tension. Cells then adhere only in the filled portion of the channel, producing a very sharp boundary. The process can be performed in an unmodified microfluidic channel with only a manual syringe and can be repeated multiple times to pattern cocultures or tricultures. We demonstrate the patterning method with two different mammalian cell types, 3T3 fibroblasts and NMuMG epithelial cells, and channel heights of 1.5 mm and 0.5 mm. We anticipate that this method will be useful for studies of cell-cell interactions where precise control of the fluidic microenvironment is required.

Tissues are increasingly being analyzed at this single cell level. We present a simple and inexpensive microfluidic filter device that can rapidly and effectively improve the quality of single cell suspensions obtained from digested tissue samples.

The endometrium is the inner lining of the uterus. Following specific cyclic hormonal stimulation, endometrial stromal fibroblasts (stroma) and vascular endothelial cells exhibit morphological and biochemical changes to support embryo implantation and regulate vascular function, respectively. Herein, we integrated a resin-based porous membrane in a dual chamber microfluidic device in polydimethylsiloxane that allows long term in vitro co-culture of human endometrial stromal and endothelial cells. This transparent, 2-μm porous membrane separates the two chambers, allows for the diffusion of small molecules and enables high resolution bright field and fluorescent imaging. Within our primary human co-culture model of stromal and endothelial cells, we simulated the temporal hormone changes occurring during an idealized 28-day menstrual cycle. We observed the successful differentiation of stroma into functional decidual cells, determined by morphology as well as biochemically as measured by increased production of prolactin. By controlling the microfluidic properties of the device, we additionally found that shear stress forces promoted cytoskeleton alignment and tight junction formation in the endothelial layer. Finally, we demonstrated that the endometrial perivascular stroma model was sustainable for up to 4 weeks, remained sensitive to steroids and is suitable for quantitative biochemical analysis. Future utilization of this device will allow the direct evaluation of paracrine and endocrine crosstalk between these two cell types as well as studies of immunological events associated with normal vs. disease-related endometrial microenvironments.

The ability to harvest single cells from tissues is currently a bottleneck for cell-based diagnostic technologies, and remains crucial in the fields of tissue engineering and regenerative medicine. Tissues are typically broken down using proteolytic digestion and various mechanical treatments, but success has been limited due to long processing times, low yield, and high manual labor burden. Here, we present a novel microfluidic device that utilizes precision fluid flows to improve the speed and efficiency of tissue digestion. The microfluidic channels were designed to apply hydrodynamic shear forces at discrete locations on tissue specimens up to 1 cm in length and 1 mm in diameter, thereby accelerating digestion through hydrodynamic shear forces and improved enzyme-tissue contact. We show using animal organs that our digestion device with hydro-mincing capabilities was superior to conventional scalpel mincing and digestion based on recovery of DNA and viable single cells. Thus, our microfluidic digestion device can eliminate or reduce the need to mince tissue samples with a scalpel, while reducing sample processing time and preserving cell viability. Another advantage is that downstream microfluidic operations could be integrated to enable advanced cell processing and analysis capabilities. We envision our novel device being used in research and clinical settings to promote single cell-based analysis technologies, as well as to isolate primary, progenitor, and stem cells for use in the fields of tissue engineering and regenerative medicine.

Micropatterned cocultures are a useful experimental tool for the study of cell-cell interactions. Patterning methods often rely on sequential seeding of different cell types or removal of a barrier separating two populations, but it is difficult to pattern sharp interfaces between pure populations with low cross-contamination when using these approaches. Patterning by the use of reconfigurable substrates can overcome these limitations, but such methods can be costly and challenging to employ in a typical biology laboratory. Here, we describe a low-cost and simple-to-use reconfigurable substrate comprised of a transparent elastic material that is partially cut to form a slit that opens when the device is stretched. The slit seals back up when released, allowing two initially separate, adherent cell populations to be brought together to form a contact interface. Fluorescent imaging of patterned cocultures demonstrates the early establishment of a sharp cellular interface. As a proof of principle, we demonstrate the use of this device to study competition at the interface of two stem cell populations.

Microfluidic devices with integrated pneumatic logic enable automated fluid handling without requiring external control instruments. These chips offer the additional advantage that they may be powered by vacuum and do not require an electricity source. This work describes a microfluidic converging-diverging (CD) nozzle optimized to generate vacuum at low input pressures, making it suitable for microfluidic applications including powering integrated pneumatic logic. It was found that efficient vacuum pressure was generated for high aspect ratios of the CD nozzle constriction (or throat) width to height and diverging angle of 3.6(o). In specific, for an inlet pressure of 42.2 psia (290.8 kPa) and a volumetric flow rate of approximately 1700 sccm, a vacuum pressure of 8.03 psia (55.3 kPa) was generated. To demonstrate the capabilities of our converging - diverging nozzle device, we connected it to a vacuum powered peristaltic pump driven by integrated pneumatic logic and obtained tunable flow rates from 0 to 130 μL/min. Finally, we demonstrate a proof of concept system for use where electricity and vacuum pressure are not readily available by powering a CD nozzle with a bicycle tire pump and pressure regulator. This system is able to produce a stable vacuum sufficient to drive pneumatic logic, and could be applied to power automated microfluidics in limited resource settings.