Date Speaker Lab
01/21 Bo Broadwater Harold Kim
01/28 Anthony and Curtis JC Gumbart
02/04 Daria Monaenkova Dan Goldman
02/11 Henry Astley Dan Goldman
02/18 Thomas Spencer David Hu
02/25 Jing Ren David Hu
03/03 Usama Sikandar Simon Sponberg
03/10 Jim Waters Harold Kim
03/17 Sunny Hwang JC Gumbart
03/31 Abouzar Kaboudian Flavio Fenton
04/07 Shane Jacobeen Peter Yunker (guest)
04/14 Diana Chen Flavio Fenton
04/21 Joey Leung Joshua Weitz
04/28 Jennifer, Will and Christian Dan Goldman
Spring 2016 Abstracts
21 January 2016
DNA strand displacement with a mismatch
Abstract: DNA strand displacement occurs between 3 DNA strands when one strand fully hybridizes to a partially duplexed complement. In biology, it occurs as
a fundamental reaction as a component of both recombination and repair of DNA. Further, strand displacement reactions are utilized as modular “tinker toys”
that combine to form complex networks with applications in biosensing, DNA locomotion, and DNA computation. In this week’s PoLS meeting, I will discuss
our work towards understanding the displacement rate as a function of single nucleotide mismatch position. Our work reveals a non-monotonic and surprisingly large change in rate due to mismatch position. Finally, I will provide a phenomenological model to quantitatively explain the salient features of the observed trend.
28 January 2016
Anthony Hazel and Curtis Balusek
The Physical Mechanisms that Drive Protein Evolution
Author: Anthony Hazel*, Curtis Balusek*, Amal Punnoose, Claire Hanson, Nick Schappaugh
Abstract: Mutations in the genome are a major driver of evolution for all organisms. Natural selection imposes “selective pressure” on these mutations, leading to differing mutation rates between species, and even between distinct regions of the genome within a single species. Protein coding sequences are of particular importance because their effects on an organism’s fitness can be directly observed if the coded protein is known. For almost a quarter century, beginning in 1975, it was assumed that a protein’s function was the primary driver of protein evolution rates. Beginning around the turn of the century, however, it was observed that a protein’s function had only a minor influence of evolution rates; rather, it was the stability of the protein’s three-dimensional structure and its ability to fold correctly which accounted for the majority of protein evolution rates. In this talk, we will discuss the Mistranslationally-Induced Misfolding (MIM) theory of protein evolution, as proposed by Drummond and Wilke in 2008, and three physical mechanisms that drive protein evolution: (1) protein folding and stability, (2) proteostasis and protein-protein interactions, and (3) translational efficiency and codon bias. Additionally, we will explore how protein folding chaperones may have reduced the selective pressure to stabilize protein folds over time.
04 February 2016
Autonomous robotic diggers provide insight into the challenges of collective nest construction
Abstract: The subterranean nests of S. invicta fire ants are stunningly sophisticated. Nest construction is accomplished through the collective effort of multiple workers. Our laboratory experiments revealed that the workload distribution within excavating groups of S. invicta was unequal. That is, a disproportionally large amount of effort is accomplished by a few workers, while the least active workers contribute to less than 1% of the collective effort. To better understand the challenges and advantages of such a workload organization we built robotic diggers. Small groups of fully autonomous robots governed by environmental clues were set to excavate simulated cohesive soil. The collective behavior within the group was coordinated by one of two social protocols. In groups governed by the first protocol all robots contributed to excavation continuously and equally. In groups governed by the second protocol the workload was asymmetrically distributed among workers. The experiments showed that with an increase in the size of the excavating group the amount of interactions between workers within the tunnel grew and led to traffic jams. The jamming slowed excavation rates while it amplified nest excavation costs. The asymmetric workload distribution resulted in the reduction of jamming. The largest tested excavating group with asymmetric workload distribution showed on average 35% higher excavation rates at 2.2 times lower excavation costs as compared to the equal workload distribution. We hypothesize that the asymmetric workload distribution could be beneficial for task performance when the resources for the task (space in our experiments) are limited.
11 February 2016
The Control and Evolution of Sidewinding
Abstract: Elongate, limbless body plans are widespread among animals, allowing movement through a wide range of spatially complex and obstacle-dense
habitats. This versatility has led to efforts to construct “snakebots”, robotic systems which mimic this morphology in hopes of achieving a similar scope of locomotor performance. However, effective limbless locomotion requires coordinating the large number of joints, whether vertebrae or servomotors, which can prove challenging. Using sidewinding locomotion as a model system, I investigate how limbless organisms and robots can simplify control while retaining behavioral versatility using a “neuromechanical template”, a simplified model of the locomotor behavior which can serve as a target for locomotor control. By independently modulating waves of vertical and horizontal movement which pass down the body of the snake, sidewinder rattlesnakes (Crotalus cerastes) can execute a variety of turning behaviors while simultaneously simplifying control. These control mechanisms were successfully applied to a sidewinding snake robot (Choset lab, CMU), dramatically improving its maneuverability while reducing control complexity. Subsequent tests in this system have yielded insights into obstacle negotiation in sidewinders, as well as the evolution of sidewinding and its distribution among snake species.
18 February 2016
Pheromone Capture in Moth Antennae
Abstract: Moths are reported to smell each other from over 7 miles away, locating each other with just 200 airborne molecules. In this study, we investigate how the structure of the antennae influences particle capture. We measure the branching patterns of over 40 moths, across two orders of magnitude in weight. We find that moth antennae have 3 levels of hierarchy, with dimensions on each level scaling with body size. Using an individual fiber mimic, we show experimentally the best orientation for particle deposition. We investigate our experimental results by performing lattice-Boltzman simulations to determine optimal antennae branch structure allowing for capture of small particles.
25 February 2016
Camouflage Feeding: Leaf Bite Patterns Are Proportional to Beetle Body Size
Abstract: on a leaf can be dangerous. The distinctive body shape and color of leaf beetles make them easily spotted by predators. In this study, we show that leaf beetles bite leaves in particular patterns to camouflage themselves. We measure the hole dimensions of 35 leaf beetle individuals. Beetles across two orders of magnitude in body mass bite holes that are 0.31 times of their body surface area. Why is hole size so consistent? We perform time-lapse photography of feeding and micro-CT scanning of the foregut. We find that hole size is constrained by two physiological factors: the mobility of the head-prothorax and the size of the foregut. We created a computer program to show that hole feeding is an active form of camouflage. The measurements from human observers showed that the more holes and the hole size closer to the body size, the better to camouflage the leaf beetles themselves. In this study, we found a new and interesting camouflage behavior and provided a new sight on the studies of insect camouflage. This master camouflage behavior is also helpful to understand why insects are the most prosperous animals in evolution.
03 March 2016
How do flight mechanics make some hawkmoths more agile than the others?
Abstract: Moths feed by hovering over swaying wind-blown flowers during day and/or night by employing immense maneuverability, agility and sensory adjustment. In an attempt to understand their amazing capabilities, we need to quantitatively characterize their overall sensory, neural and body dynamical system. A hawkmoth Manduca sexta feeds during twilight while adjusting well to dim light by increasing its visual exposure time. But two species of hawkmoth are specialized with respect to light conditions: Macroglossum stellatarum feeds during the day while Deilephila elpenor is truly nocturnal. Deilephila’s bright light response can be constructed by increasing the exposure time and visual sensitivity of Macroglossum’s, but there is no way we can construct Manduca’s response by varying just these two parameters. Also, the Deilephila performs the worst at following the swaying flowers. Additionally, we know that Macroglossum is on average 2-3 times smaller than Manduca but flaps its wings around thrice the Manduca’s wingbeat frequency, while Deilephila’s stands as an intermediate. Therefore, mechanics may be able to explain the performance differences mentioned above. We constructed a linear time-averaged dynamical model to define the flight mechanics of each hawkmoth species and analyzed the eigenmodes to find out the variation in each mode’s stability across the three species. We found that forward speed coupled with pitching rate is unstable for all. However, on a wing-beat-scaled time, the roll coupled with yaw is found to be much more stable in Manduca than Macroglossum, while Deilephila stays as an intermediate. On the other hand, Macroglossum’s flight appears to be more maneuverable and stable than Deilephila’s. This could either mean that Macroglossum’s neural feedback control does much better job than Deilephila’s, or that the stability vs maneuverability tradeoff just does not apply in this case.
10 March 2016
Predicting DNA Unlooping with Phase-Space Sampling
Abstract: DNA-binding proteins can regulate genetic expression by holding two sites in close proximity, forming a closed loop. Such complexes may require strong bending of DNA segments on the order of one persistence length or less. Both this elastic bending and the thermal fluctuations of the DNA molecule are necessary to describe the resulting behavior. To explore this problem, we consider a discrete model of a wormlike chain, kept in the fixed extension ensemble. By using a novel method to sample conformations in both position and momentum space, we can obtain a distribution of constraint forces as a function of chain length, extension, and flexibility. Our coarse-grained model allows us to explore the space of these parameters more efficiently than a detailed molecular dynamics approach. We find that increasing contour length decreases average force by relieving bending stress, but that the additional freedom allows fluctuations in the constraint force to increase. This implies that the probability of large forces may go up even as the mean goes down, impacting the lifetime of such bound states in a way unforeseen by purely equilibrium methods.
17 March 2016
Determination of the mechanical properties of realistic bacterial inner and outer membranes
Abstract: The cell envelope in Gram-negative bacteria is made of two distinct membranes and a cell wall between them. From a mechanics point of view, the
cell maintains a higher concentration of solutes in the cytoplasm than in the external environment and the difference in osmotic pressure, known as the turgor pressure, places a great stress on the envelope, although which element bears the most is unclear. Now, we have used molecular dynamics (MD) simulations of membranes to resolve how lipid membranes respond to changes in lateral tension resulting from the turgor pressure. In this study, we used two models of the inner, cytoplasmic membrane; the first membrane is modeled as a mixed 75%POPE/25% POPG bilayer while the second membrane model consists of saturated, unsaturated, and cycle-containing lipids that more accurately reflect the diverse population of lipids within the E. coli cytoplasmic membrane. Additionally, we looked at the bacterial outer membrane, which has an outer leaflet of lipopolysaccharides that stiffen it. We applied surface tensions of values from 10 to 100 dynes/cm and measured a variety of properties of membranes (area per lipid, thickness, etc.), thus providing a quantitative description of the membrane response. More general mechanical properties of membranes were also characterized, namely the elastic area compressibility modulus and Young’s modulus, in order to describe the elasticity of membrane. Our work demonstrates that differences in lipid composition result in a differential response to lateral tension.
31 March 2016
GPU/WebGL: A Pathway to Real Time Patient Specific Heart Modeling
Abstract: According to the Center for Disease Control (CDC), more than 600,000 Americans die of heart disease each year, which means that one in every four deaths is caused by heart disease. Patient specific simulations not only will
lead to better understanding of arrhythmia initiation but also design of better control and prevention strategies. However high resolution simulations in anatomically accurate 3D structures using realistic cardiac cell models for
a couple of seconds requires solving on the order of a trillion differential equations, thus to date only achieved using supercomputers. However, the development of Graphical Processing Units (GPUs), in modern computers and even
cell phones, gives them the amazing power of high performance computing. Effectively, each of our GPU enabled devices, can become a super computer. In this work, we will show how WebGL can be used to solve complex cardiac
problems in a high-performance framework, inside a web-browser and independent of the operating system. This approach makes our parallel solver extremely fast, and easily accessible to the end users. We will show how this method,
can enable us to perform near real time, interacting modeling of cardiac dynamics in two and three-dimensions. This is a big first step towards a viable and affordable approach into patient specific modeling.
07 April 2016
The Role of Physics in the Evolution of Multicellular Complexity
Abstract: The evolution of early multicellularity as much about physics as it is about biology, as selection acts on the physical properties of multicellular bodies. Nascent multicellular organisms are confronted by new internal and external forces that act on length scales larger than single cells and are capable of fracturing intercellular bonds. We study the
evolution of the mechanical properties of multicellular ‘snowflake’ yeast that were selected for increased size over ~1,500 generations. While these snowflakes evolve to be larger by mitigating internal forces, they also become more susceptible to fracturing when faced with external compressive forces. Using confocal microscopy and direct mechanical measurements, we investigate the physical underpinnings and consequences of this strength-toughness trade-
14 April 2016
Diana Chen (Diandian)
A Mechanism for QRS Alternans and the Initiation of Spatiotemporal Chaos
Abstract: Cardiac arrhythmias cause thousands of deaths per year. The deadliest arrhythmia is Ventricular Fibrillation, which can be caused by reentry of an electrical signal. This reentry can result in multiple spiral waves that lead to the uncoordinated contraction of the heart, which will result in death within minutes if left untreated. The causes of reentry can stem from cellular level dynamics. A common method for detecting instability is action potentional duration (APD) alternation of a cellular electrical signal from cardiomyocytes. However, clinically it has been shown that Action potential amplitude alternation correlates more strongly with ventricular fibrillation. In addition, in a majority of simulations and experimental data, it has been shown that though APD can alternate greatly from beat to beat, no conduction block will occur without Action potential amplitude alternation. Thus, we proceed to describe a new mechanism for cellular electrical instability due to APA alternation by deriving a map and a mathematical model. The map gives us excellent predictions of instability at the cellular level and the cell model in tissue leads to break up spiral waves due to APA alternation.
21 April 2016
Modeling the dynamics of bacteria, phage, and the immune system: when does phage therapy work?
Abstract: The rise of pervasive antibiotic resistance has led to a renewal of interest in using bacteriophage (phage) to treat bacterial infections . Phage therapy has been viewed as a potential treatment for over a century. Yet this year marks the first phase I/II human trial of a phage therapeutic – to treat burn wound patients in Europe. The slow progress in realizing clinical therapeutics is matched by a similar dearth in theoretical understanding of how, why, and when phage therapy works. In contrast to the stated goals of phage therapy, standard phage-bacteria models and in vitro experiments often lead to coexistence of phage and bacteria. Instead, it has been hypothesized that phage together with an immune system can synergistically eliminate a bacterial pathogen. Existing models of bacteria-phage-immune system dynamics include simplified assumptions such as unbounded growth of the immune response or the lack of immune evasion by bacteria [2, 3]. Moreover, they do not provide a mechanistic basis for synergistic effectiveness of phage therapy. Here, we propose a model of phage therapy that incorporates a maximum capacity of the immune response and immune evasion by bacteria at high density. We identify a synergistic regime in which phage and the immune response jointly contribute to the elimination of the bacteria. Crucially, we find that in this regime, neither phage alone nor the immune system alone can eliminate the bacteria. We study the mechanism underlying the synergistic effect and its significance for different infection and immune parameters of potential clinical relevance.
 R. Young and J. J. Gill, Science 350, 1163 (2015).
 B. R. Levin and J. J. Bull, Nature Rev. Microbiol. 2, 166 (2004).
 K. Hodyra-Stefaniak et al., Sci. Rep. 5, 14802 (2015).
21 April 2016
Christian Hubicki, Jennifer Rieser, and Will Savoie
Planning Dynamic Robot Locomotion on Granular Media: An Impulsive Jumping Study
Abstract: This work demonstrates fast motion planning for robot locomotion that is optimized for terrain with complex dynamics, specifically, rapid penetration of granular media. Gait planning is critical for many legged locomotion control approaches, but they typically assume rigid ground contact. We aim to extend these planning methods to include terrain dynamics we see in the natural world, like sand and dirt, which can both deform and fluidize. Using an added-mass dynamical description of collective grain motion, we were able to formulate a description of hydrostatic and hydrodynamic terrain effects that is both principled and representable with closed-form dynamics. As a result, we present a model and fast optimization formulation which solves accurate motion plans on granular media with tractable solving times (between 1-10 seconds). For validation, we optimized open-loop motor trajectories for a testbed jumping robot to jump to a target apex height from a bed a loosely packed poppy seeds. While jumps optimized for rigid ground were anemic on granular media, terrain-aware trajectories consistently hit within 4mm of their target. This demonstrates the potential for designing accurate robot locomotion trajectories which respect practical task specifications, all while being aware of the terrain beneath it.
Scattering of a limbless locomotor in complex terrain
Abstract: Animals can often navigate through complex terrain with seemingly little effort. Despite this, numerous robotic attempts to mimic animal behavior in even simple situations have shown that the interactions between animals and their environment are far from simple. To gain some insight into complex animal-environment interactions, we focus on a limbless locomotor in a simplified environment: a snake traversing a one-dimensional array of evenly-spaced pegs. Both with a robotic snake and simulations, we find that the snake trajectory is altered by the interactions with the peg array, and that there are strongly preferred final trajectories. We explore how these preferred directions of travel depend on the geometry of both the peg array and the snake shape.
Smarticles: smart, active granular matter
Abstract: We investigate a granular medium composed of smart, active particles, or “smarticles”. Previously, we discovered that ensembles of “u”-shaped particles exhibited geometrically-induced cohesion by mechanically entangling via particle interpenetration [Gravish et al, PRL, 2012]; the strength and/or extent of entanglement could be varied by changing particle level entanglement by changes in arm-to-base length of the u-particle. Since changing this parameter on demand is inconvenient, we develop a power-autonomous programmable robot composed of two motors and three links with an on-board microcontroller. This smarticle can be activated to change its configuration (specified by its two joint angles) through audio communication. To complement these experiments, since study large ensembles of smarticles is cost and labor prohibitive, we also develop a simulated smarticle in the Chrono multibody simulation environment. We systematically study ensemble cohesiveness and compaction as a function of shape changes of the smarticles. We find that suitable activation of smarticles allows ensembles to become cohesive to “grip” rigid objects and lose cohesion to release on command.