Human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CM) have great potentials in biomedical research and can be used extensively in drug screening and heart simulations. To understand the cardiomyocytes, we need to perform functional analysis on these muscle cells. Therefore, we need a simple, controllable, yet biocompatible and high throughput tool to measure the cellular traction force. At the Sniadecki Lab, we are developing a new technique to measure the force generation of hiPSC-CM: dotted traction force microscopy platform. To create the platform, fluorescent proteins were first absorbed to a dotted polydimethylsiloxane (PDMS) negative and stamped onto a polyvinyl alcohol film. The film was then transferred to a soft PDMS substrate and subsequently dissolved using phosphate buffered saline solution while the patterned fluorescent proteins stained the substrate. Since the stiffness of the soft PDMS substrate is known, the force generation of the cardiomyocytes can be calculated in real time by optically tracking the deformation of the fluorescent dots. Currently, we are able to manufacture the platform with high fidelity and uniform alignment with a production time of less than 2 hours. Moreover, the cardiomyocytes can fully spread out to their in vivo state on the substrate which ensures the force measurement is valid and accurate. Potentially, this method is not limited to cardiomyocyte research and can be applied to study the interaction between force generation and cell performance of other cells. We are also exploring the possibility of automated manufacture and integration with 96-well to enable mass production.

In computer graphics, a voxel (volume element) is a point in a 3D world coordinate system (i.e. the coordinate system of a virtual world). In games like Toca Blocks or Minecraft, voxels are used to store the texture of a particular terrain. Sometimes, voxels next to each other have the same texture. When voxels of homogeneous textures form polygons, rendering systems will optimize memory storage by storing the polygons' vertices rather than every single voxel in the polygon. The process of choosing polygons that cover the voxels is known as meshing. We refer to these polygons as quads and the collection of quads as a mesh. Current methods for polygon meshing require too much data storage or require a drastic change in the mesh after a small change in the world coordinate system. We propose the Greedy Face Meshing (GFM) Algorithm, a linear time algorithm for meshing voxels into quads. We prove that our algorithm is within a constant factor of the optimal solution (in terms of number of quads) and can update in constant time for a single-voxel change in the world coordinate system. We also show how the GFM Algorithm can be implemented using the Segment Tree data structure. Rendering systems can use the GFM algorithm to mesh polygons since its storage is no worse than any existing algorithm and its updates take constant time.

Rats produce ultrasonic vocalizations (USVs) in a range from 20-kHz to 95-kHz that vary in frequency and shape across social and motivational contexts and can correspond to the affective state of the animal. To assess these USVs accurately and efficiently, our lab created DeepSqueak, a novel machine learning software package that expedites the detection and analysis of rat USVs by using neural networks to differentiate them from noise. DeepSqueak also allows for automatic and unbiased classification of USVs into discrete categories using call parameters such as shape, frequency and duration. Prior to this unbiased categorization method, identified 14 subjective categories in 50-kHz rat vocalizations that could be manually identified by a trained experimenter. These categories have received some limited study, but the excessive labor and time needed for manual classification restricted broad adoption. We aim to use neural networks to quickly and automatically classify USVs into these categories to promote broad adoption and better our understanding of the relationship between USVs and behavior. The process of training our neural network to differentiate between vocalizations was approached in two ways. Audio files were converted to sonograms through DeepSqueak and manually labeled. Thousands of these labeled calls were then inputted as training data for the neural network. This method allowed the network to learn using a large set of labeled vocalization data. The second method is based around the manual selection of an optimal call for each subtype using DeepSqueak's "call clusters" function; the neural network was then trained around how closely vocalizations matched the optimal calls. We now plan to compare DeepSqueak's automated calls and clustering to manual scoring in order to develop the best possible system that reliably categorizes USVs, thus allowing for more specific analyses of USV categories and behavior.

Chlamydia trachomatis is a human urogenital pathogen and the leading cause of bacterial sexually transmitted infection worldwide.  A major aim of the Hybiske Lab is to develop a functional genetic understanding for Chlamydia. I am studying a set of newly generated C. trachomatis chimeric mutants that were generated from interspecies lateral gene transfer between C. trachomatis and the mouse adapted species C. muridarum. This series of recombinant strains contain a differing extent of genetic exchange surrounding the predicted inclusion membrane protein (Inc) CT147. CT147 is predicted to be secreted into the Chlamydia-containing vacuole (inclusion) membrane by type III secretion and subsequently mediate molecular interactions with host proteins. In cultured cells, strains lacking CT147 prematurely rupture inclusions at 24 hours post infection, in stark contrast to wildtype C. trachomatis or control recombinant strains that grow normally inside host cells and do not exhibit inclusion lysis at any stage of infection. We therefore hypothesized that the C. muridarum ortholog of CT147 is incompatible with the series of ~30 Inc proteins normally secreted by C. trachomatis, in such a way that inclusion integrity is not properly maintained during this strain’s developmental growth. I have used quantitative RT-PCR to evaluate transcript levels early and late in each strain’s life cycle to determine if the diversion in phenotypes is due to an alteration in the Inc’s promoter. I have expressed in trans CT147 in the chimeric strain lacking the Inc and analyze if CT147 can rescue the premature rupturing phenotype. Similarly, I expressed the C. muridarum ortholog TC0424 into wild type C. trachomatis to determine if the phenotype is dominant. Overall, we anticipate that a molecular and functional characterization of this novel Inc protein will reveal new insight into the mechanisms by which Chlamydia manipulate host cell function to facilitate their infection.