Fall 2016 Schedule

Date Speaker Lab
09/01/16  Mike Tennenbaum  Fernandez-Nieves Lab
09/08/16 Zachary Jackson, and Nate Conn  Wiesenfeld Lab and Goldman Lab
09/15/16  Yasemin Ozkan Aydin  Goldman Lab
09/22/16 Ben Kalziqi and Conner Herndon  Yunker Lab and Fenton Lab
09/29/16  Yanyan Ji and Neil Hardy  Fenton Lab
10/06/16 Izaak Neveln Sponberg Lab
10/13/16 Karl Lundquist Gumbart Lab
10/20/16 Ashley Coenen and Yu-Hui  Lin Weitz Lab
10/27/16 Jianing Wu Hu Lab
11/03/16 Perrin Schiebel, Alex Hubbard, and Jennifer Rieser Goldman Lab
11/10/16 Christian Hubicki Goldman Lab
11/17/16 Jiyoun Jeong Kim Lab
12/01/16 Olga Shishkov Hu Lab

Fall 2016 Abstracts


Mike Tennenbaum (Fernandez-Nieves Lab)

Reconfigurable mechanics of fire ant aggregations

Fire ant aggregations are an inherently active system. Each ant has its own energy source and can convert this energy into motion. We find that the properties of ant aggregations changes cyclically in time. These cycles are connected to the activity level of the aggregation. We monitor the mechanics by measuring the normal force, oscillatory rheology, and real space imaging. With these measurements we can connect low activity levels to density heterogeneities, more elastic mechanics, and low normal force while high activity is connected to more homogeneous density, power law rheology, and higher normal force. We interpret out results with a model that considers active and inactive ants.


Zachary Jackson (Wiesenfeld Lab)

Hair Cells as a Source for Tinnitus

Tinnitus, or ringing in the ear, affects around 10% of the US population. Hair cells are a source of nonlinear amplification in the ear and could be a cause of certain types of the disorder. When their stereocilia are deformed they quickly push against the basilar membrane. Ongoing work is presented on a mathematical model of hair cells as coupled oscillators. Under this model, sustained oscillations are demonstrated in the absence of external stimulation. Also presented is a healing parameter and a noninvasive therapy for tinnitus sufferers.

Nate Conn (Goldman Lab)

The kinematics of root growth during interaction with obstacles in gel

As roots grow, they must navigate highly complex granular environments to anchor and retrieve water and nutrients. From gravity sensing at the root tip to pressure sensing along the tip and elongation zone, the complex mechanosensory feedback system of the root allows it to bend towards greater depths and avoid obstacles of high impedance by asymmetrically suppressing cell elongation. Here we investigate the various mechanical and physiological responses of roots to mechanical stimuli. We grow Maize plants in 2D glass containers filled with photoelastic gel and observe root and shoot growth. Our automated system collects photo timelapses of multiple plants. Interaction with a variety of different shaped obstacles seem to coincide with changes in root/shoot growth. For example, analysis of root curvature in some plants resulted in a corresponding change in the curvature of the shoot. We hypothesize that this may be attributed to a combination of signalling between the root and shoot as well as a mechanical response.


Yasemin Ozkan Aydin (Goldman Lab)

Coordinated Body Bending Improves Performance of a Salamander-like Robot on Granular Medium

Mudskippers and salamanders can both swim and navigate outside of the water. We wish to understand how limbs and body morphology contribute to performance, particularly in the evolution of animals that use multiple modes of locomotion. Previous study investigated the tail use in the earliest land animals, and using robophyscial model it was shown that properly coordinated tail movement helps to propel body weight. Here we study a fire salamander (S. salamandra) to understand the coordinated movement between the flexible body and front/back limbs. Experiments (10 adults, 5 trials each, varying inclines and presence/absence of sandy surface) revealed that salamanders propelled themselves using diagonal leg pairs and body undulation. To investigate mechanisms governing effective locomotion, we built a robophysical model and tested its performance on hard ground (HG) and yielding granular media (GM) of poppy seeds of different inclinations. Our servo-driven robot (430 g, 38 cm long) has four limbs, a flexible trunk, and an active tail. Each limb has two motors to control vertical position and step size of limb. A joint in the middle of the body controls horizontal bending. We assessed performance of the robot by measuring body displacement over a few limb cycles. On HG at 0o and 10oslopes, the robot performed well; feet did not slip and body bending increased step length (SL) by ~50% (on level) and ~1% (on 10o slopes ). On GM, the robot’s SL decreased by ~40% (on level) and ~80% (on 10o slopes) relative to that on HG due to limb slip. Back bending restored performance on GM, leading to SLs only ~20% (on level) and ~25% (on inclined) less than those on HG. A geometric mechanics model revealed that on level GM media body bending was most beneficial when phase offset 180 degrees from leg movements.


Ben Kalziqi (Yunker Lab)

Killing to Equilibrate

Unlike equilibrium atomic solids, densely packed tissues and biofilms do not experience significant thermal fluctuations at the cell level.  However, unlike atoms, cells can die and reproduce; these number fluctuations stochastically perturb cellular solids, and have been predicted (by Risler et al., PRL 2015) to produce an effective fluctuation-response relationship.. We investigate how life-and-death events affect surface-height fluctuations of biofilms with two strains of Vibrio cholera. The two strains used are mutual predators that kill on contact via the Type VI Secretion System. Biofilm surface fluctuations are measured with a white light interferometer in the homeostatic limit, wherein division and death occur at roughly the same rate. Although these processes result in a decidedly out-of-equilibrium system, the measured height correlation function lines up with expectations from a generalized fluctuation-response relation due to life and death events, predicted by Risler, et al. The resultant effective temperature increases with the amount of intercellular killing, allowing us to tune effective temperature by varying the strain number ratio and (actual) temperature.

Conner Herndon (Fenton Lab)

Emergent phenomena in paced cardiac tissue: forging a link from cellular nonlinear dynamics to electrocardiographic risk assessment.

The heart is an excitable system through which electrical waves of depolarization propagate in a coordinated manner to initiate mechanical contraction. By tracing the the extracardiac voltage over time, the electrocardiogram (ECG) can quickly identify anomalous electrical pathways at the global level, and through experimental techniques such as optical mapping, we may directly observe cellular dynamics concurrently with ECG measurement. In this talk I will discuss the experimental methods used in our lab to study cardiac electrical function from cellular to whole-heart level and how nonlinear analysis allows for a physical description and prediction of medical risk.


Yanyan Ji (Fenton Lab)

Synchronization as a Mechanism for Low-energy Anti-fibrillation Pacing (LEAP)

Background–Cardiovascular disease claims more lives than any other major cause of death in the United States. Standard defibrillation by strong shocks still remains the mainstay treatment for symptomatic patients even though strong electrical shocks have been observed to cause severe side effects. Recently, low-energy anti-fibrillation pacing (LEAP) has been suggested as an alternative treatment that significantly lowers the energy required when using multiple shocks. However, the mechanism by which LEAP terminates arrhythmia remains unclear and the correlation between the timing of the shocks and the success rate is controversial. In this study, we investigated the mechanism for arrhythmia termination by LEAP using both experiments and simulations and systematically explored the influence of shock period and timing on the success rate of LEAP.

Methods and Results–Both in vitro experiments on canine hearts and simulations showed successful termination with energy reduction for terminating atrial fibrillation and ventricular fibrillation using LEAP, and verified synchronization from virtual electrodes is the key mechanism for termination of arrhythmia by LEAP. We also observed in simulations that LEAP is more effective when the shock period is close to the dominant period and the first shock is delivered when the fraction of tissue excited is decreasing.

Conclusions–Our results support synchronization as the mechanism for effective arrhythmia termination by LEAP, and its effectiveness can be improved by adjusting shock period and timing.

Neil Hardy (Fenton Lab)

Optimal Pulse Configuration Design For Heart Stimulation. A Theoretical, Numerical, and Experimental Study.

Rationale: Existing pacemakers consider the rectangular pulse to be the optimal form of stimulation current. However, other waveforms for the use of pacemakers have not yet been tested. Time decaying exponential pulses are known to be the preferred waveform for defibrillators use, but why the exponential waveform is optimal is still unknown.

Objective: We aim to find the optimal waveform for pacemaker use, and to offer a mechanical explanation for its advantage. Since the pacemaker battery is a charge source, here we probe the stimulation current waveforms with respect to the total charge delivery.

Methods and Results: To test our theory we performed experimental studies on cat and rabbit hearts and simulations implemented on a 1D human ventricular cable. The results show that the tissue could be successfully paced both by the rectangular and by the exponential. Nevertheless, the time decaying exponential stimulations contain a lower extent of total charge, i.e., higher efficiency, in comparison to the rectangular pulses. To explain this theoretically, we combine optimal control theories and electrophysiology analyses.

Conclusions: Our theoretical analysis showed that the myocyte ion-channel currents act as an additional source of charge that adds to the external stimulating charge for stimulation purposes. Therefore, as the action potential emerges, the external stimulating current can be reduced accordingly exponentially. From the experiments, we calculated that the longevity of the pacemaker battery is ten times higher for the exponential current compared to the rectangular waveform.

Key Words: battery longevity, myocyte ion-channels currents, optimal control, optimal waveform, pacemaker.


Izaak Neveln (Sponberg Lab)

Changes in Centralization of Control of Movement as Speed Varies

Legged locomotion requires control of numerous complex coupled systems, yet strategies employed by animals such as cockroaches allow for robust navigation through varied environments. As animals move faster, one might expect that the need for coordination becomes more important but that centralizing control is made more difficult due to limiting bandwidth. Here we expand on previous work on using mutual information (MI) as a metric for centralization to test the hypothesis that faster locomotion leads to a more centralized control architecture. In the cockroach Blaberus discoidalis, we estimate the MI between leg control signals and two kinematic state variables: a local measure of leg extension and a global measure of the overall limb kinematics. MI informs how much the control signal reduces the possible variability in the output variable and vice versa without assumptions of a particular model. We section a large data set of cockroaches running on flat terrain by stride frequency. The degree of centralization, as quantified by the amount of global MI minus the local MI, peaks at an intermediate stride frequency that corresponds to the previously published preferred running speed. Therefore, while centralization may have increasing benefits with speed, there is evidence that coordinating control becomes less attainable due to bandwidth constraints.


Karl Lundquist (Gumbart Lab)

Simulations of BamA in a native outer-membrane model and energetics of lateral openings

Gram-negative bacteria possess two cell membranes. The outer membrane is host to almost exclusively beta-barrel transmembrane proteins. The beta-barrel assembly machinery  (BAM) is a five-protein complex responsible for the insertion and assembly of outer membrane beta-barrel proteins. The first crystal structures of the transmembrane beta-barrel domain of BamA were released in 2013. In these initial crystal structures, the BamA beta-barrel possessed a weak interface between its first (1) and last (16) beta strands, which led to strand separation within 1us in our equilibrium simulations. This strand separation has been proposed to act as a lateral gate for substrate passage into the membrane. In addition, a hydrophobic mismatch near the barrel seam was shown to destabilize the membrane, potentially acting to aid integration of the substrate. Furthermore, full BAM complex crystal structures were recently released, showing accessory proteins rallying around this putative insertion region using the periplasmic domains of BamA as a scaffold. In order to address remaining questions surrounding the role that BamA plays in the insertion and assembly process, we have carried out additional equilibrium simulations of BamA in several membrane bilayers. We also performed a calculation of energetic landscapes associated with  lateral gate formation under various conditions. These calculations reveal a lower energetic barrier to strand separation for BamA as compared to FhaC, a member of the same (Omp85) family. They also reveal a significant role played by the C-terminal kink feature in modulating the barrier to strand separation. Finally, equilibrium simulations of BamA demonstrate lateral gate opening in a native lipopolysaccharide bilayer for the first time.


Yu-Hui Lin (Weitz Lab)

Interplay between Biological Process and Spatial Patterns in Host-Virus Systems

This project is aimed to investigate how the emergence of spatial patterns could affect the interaction between host and virus. It has been reported that population stochasticity due to random birth events can induce clustering of individuals and thus spatial inhomogeneity within population. Local clustering of individuals increases the possibility for biological process (birth, predation, etc.) to happen in its neighborhood and could in turn change the population dynamics.  Here, we ask how do the patterns change in response to varying the number of progeny per birthing event varies. These bursty growth dynamics are particularly relevant to viruses which can release between 10 and 10000 virus particles per infection. These perturbations can not be accounted for in mean-field ODE models of predator prey.  We develop individual-based spatial models to investigate the dynamics. We then consider 3 regimes of viral growth: uninhibited exponential growth, limited logistic growth, and explicit host-viral dynamics. In the first two cases, only the viruses are modeled explicitly, and hosts are assumed to be everywhere and replenishing. In the third case, both hosts and viruses are modeled explicitly. Our preliminary result shows that increasing burst size of individuals does increase the density and size of clusters for all regimes we tested. Including limitation of logistic growth on bursting individuals lower the density of clusters and alters the equilibrium abundance predicted by ODE. The future goal is to focus more on how the emergence of the spatial patterns affect the way host and virus interact with each other, and to identify the viral burst size that leads to the most stable host-virus dynamics.

Ashley Coenen (Weitz Lab)

Who infects whom? Inferring interactions in marine virus-microbe communities.

Microbes are abundant and diverse in the ocean. The same is true for viruses of microbes. Viruses are estimated to turn over 10 to 40 percent of microbes daily. Consequently, viruses are important influences in shaping microbial communities. The interactions among microbes and viruses (i.e. “who infects whom”) are difficult to pin down in situ. Through a variety of techniques, we can obtain population time series of virus-microbe communities. Deducing which pairs interact from time series (the “inference problem”) is an open question in viral ecology and other fields. Here, we examine one popular approach to the inference problem, which uses correlations to infer interactions. We argue that correlations 1) do not have a clear biological interpretation and 2) cannot capture nonlinear relationships over complex networks. We test the correlation approach on a simulated virus-microbe community and introduce an alternative model-based approach.


Jianing Wu (Hu Lab)

Matter transport by elephant trunks

The boneless elephant trunk is the elephant’s most versatile appendage, enabling it to grab objects ranging from peanuts to logs, and siphon water or mud in the open air. We discover that elephants can be surprisingly gentle. Specifically, they push down on objects with only 5 percent of their trunk weight, less than half the weight that blindfolded humans apply to objects. This ability is made possible by the elephant forming joints with its trunk. The joint separates the trunk into two sections, the latter of which pushes down on objects purely with its self-weight. Elephants change the location of the joint accordingly to modify the downward force necessary to gather the food particles together. The elephant can siphon water at a high volumetric rate of 5 L/s, which . We measured the siphoning rate of an elephant’s siphoning water and discovered that, the elephant may have a potential of expanding air passages to augment the water intake rate. This work may inspire ways to control soft robotic actuators and design new types of liquid transporters.


Perrin Schiebel, Alex Hubbard, and Jennifer Rieser (Goldman Lab)

A robophysical model for limbless locomotion in a heterogeneous environment

Snakes move gracefully through varied terrain, negotiating obstacles such as twigs, rocks, and grasses. Despite the seeming simplicity of this movement, the continuous interaction with the ground coupled with obstacle collisions can give rise to complex dynamics. Work in our lab that explores the interaction of the desert-dwelling sand specialist C. occipitalis with a row of vertical pegs (perpendicular to the initial direction motion) has found that the snake is least likely to apply forces to the pegs along the direction of motion. We have also observed that the snake does not substantially alter its waveform to maneuver through pegs, suggesting that positional control of shape is a reasonable neuromechanical control model. To test this, we built a 13 segment servo-motor-driven snake-like robot (1.13 kg, 80 cm long). Joint angles were commanded via the motors, and low slip translational motion of the robot was achieved by affixing wheels to each segment. To sample robot-peg interactions, the initial shape of the robot was fixed and the robot was placed at different locations within a rectangular region (dimensions set by the peg spacing and distance traveled over one period). The robot position was recorded over several cycles, and in-plane reaction forces were measured via strain gauges on each peg. The forces were complex, with multiple collisions occurring during each transit, but a simple pattern emerged: the distribution of the force orientations for the robot was similar to that of the snakes. This suggests that the animal’s interaction with obstacles is dominated by its sand-adapted body wave control mechanics. Intriguingly, video tracking of the robot after transit revealed that it was re-oriented along preferred directions, as might be predicted by wave mechanics.


Christian Hubicki (Goldman Lab)

Objectives of Bipedal Locomotion from Cursorial Birds to Robots

Bipedal locomotion is a complex phenomenon to understand and control, making it difficult for legged robots to achieve the speed, efficiency, and robustness of their animal counterparts. Here we investigate the underlying task-level control objectives of dynamic locomotion in ground-running birds, with an eye toward application on bipedal robots. Specifically, we seek to tease out how birds balance demands on stability and energy economy while traversing uneven terrain. This work uses trajectory optimization as its primary tool for analyzing bipedal locomotion data from experiments with obstacle-traversing birds, ranging in size from quail to ostrich. The results suggest that dynamic and biologically comparable bipedal locomotion can be achieved by minimizing energy costs while strictly avoiding injurious forces and satisfying practical locomotion task constraints. We further present experiments using the bipedal robot, ATRIAS, as a step toward validating these concepts on robotic hardware.


Jiyoun Jeong (JJ) (Kim Lab)

Observation of flexibility reversal in DNA bending

Several experiments on DNA looping show that DNA looping probability at short length scales is higher than the prediction of the wormlike chain model. This observation suggests that DNA becomes more flexible at large bending angles and hints at the possibility that sequence dependence of flexibility may not be universal across different bending regimes. Using a FRET-based DNA looping assay, we measured flexibilities of various DNA sequences (some of which include base pair mismatches) from the looping and unlooping rates. Surprisingly, we find a strong correlation between the measured looping and unlooping rates, which points to an apparent flexibility reversal: more flexible sequences in the unlooped state are more rigid in the looped state. To explain this counterintuitive finding, we present a few hypotheses that challenge our coarse-grained level understanding of DNA.


Olga Shishkov (Hu Lab)

Active mixing increases feeding rate of black soldier fly larvae

How do we sustainably feed a growing world population? One solution of increasing interest is the use of black solider fly larvae, pea-sized grubs envisioned to transform hundreds of tons of food waste into a sustainable protein source. Although startups across the world are raising these larvae, a physical understanding of how they should be raised and fed remains missing. In this study, we present experiments measuring their feeding rate as a function of number of larvae. We show that larger groups of larvae have greater mixing which entrains hungry larvae around the food, increasing feeding rate. Feeding of larvae thus differs from feeding of cattle or other livestock which exhibit less self-mixing.