Revealing principles of autonomous thermal soaring in windy conditions using vulture-inspired deep reinforcement-learning.

Thermal soaring, a technique used by birds and gliders to utilize updrafts of hot air, is an appealing model-problem for studying motion control and how it is learned by animals and engineered autonomous systems. Thermal soaring has rich dynamics and nontrivial constraints, yet it uses few control parameters and is becoming experimentally accessible. Following recent developments in applying reinforcement learning methods for training deep neural-network (deep-RL) models to soar autonomously both in simulation and real gliders, here we develop a simulation-based deep-RL system to study the learning process of thermal soaring. We find that this process has learning bottlenecks, we define a new efficiency metric and use it to characterize learning robustness, we compare the learned policy to data from soaring vultures, and find that the neurons of the trained network divide into function clusters that evolve during learning. These results pose thermal soaring as a rich yet tractable model-problem for the learning of motion control.

A hull reconstruction-reprojection method for pose estimation of free-flying fruit flies.

Understanding the mechanisms of insect flight requires high-quality data of free-flight kinematics, e.g. for comparative studies or genetic screens. Although recent improvements in high-speed videography allow us to acquire large amounts of free-flight data, a significant bottleneck is automatically extracting accurate body and wing kinematics. Here, we present an experimental system and a hull reconstruction-reprojection algorithm for measuring the flight kinematics of fruit flies. The experimental system can automatically record hundreds of flight events per day. Our algorithm resolves a significant portion of the occlusions in this system by a reconstruction-reprojection scheme that integrates information from all cameras. Wing and body kinematics, including wing deformation, are then extracted from the hulls of the wing boundaries and body. This model-free method is fully automatic, accurate and open source, and can be readily adjusted for different camera configurations or insect species.

Solving the thoracic inverse problem in the fruit fly.

In many insect species, the thoracic exoskeletal structure plays a crucial role in enabling flight. In the dipteran indirect flight mechanism, thoracic cuticle acts as a transmission link between the flight muscles and the wings, and is thought to act as an elastic modulator: improving flight motor efficiency thorough linear or nonlinear resonance. But peering closely into the drivetrain of tiny insects is experimentally difficult, and the nature of this elastic modulation is unclear. Here, we present a new inverse-problem methodology to surmount this difficulty. In a data synthesis process, we integrate literature-reported rigid-wing aerodynamic and musculoskeletal data into a planar oscillator model for the fruit flyDrosophila melanogaster, and use this integrated data to identify several surprising properties of the fly's thorax. We find that fruit flies likely have an energetic need for motor resonance: absolute power savings due to motor elasticity range from 0%-30% across literature-reported datasets, averaging 16%. However, in all cases, the intrinsic high effective stiffness of the active asynchronous flight muscles accounts for all elastic energy storage required by the wingbeat. TheD. melanogasterflight motor should be considered as a system in which the wings are resonant with the elastic effects of the motor's asynchronous musculature, and not with the elastic effects of the thoracic exoskeleton. We discover also thatD. melanogasterwingbeat kinematics show subtle adaptions that ensure that wingbeat load requirements match muscular forcing. Together, these newly-identified properties suggest a novel conceptual model of the fruit fly's flight motor: a structure that is resonant due to muscular elasticity, and is thereby intensely concerned with ensuring that the primary flight muscles are operating efficiently. Our inverse-problem methodology sheds new light on the complex behaviour of these tiny flight motors, and provides avenues for further studies in a range of other insect species.

Model-Based Tracking of Fruit Flies in Free Flight.

Insect flight is a complex interdisciplinary phenomenon. Understanding its multiple aspects, such as flight control, sensory integration, physiology and genetics, often requires the analysis of large amounts of free flight kinematic data. Yet, one of the main bottlenecks in this field is automatically and accurately extracting such data from multi-view videos. Here, we present a model-based method for the pose estimation of free-flying fruit flies from multi-view high-speed videos. To obtain a faithful representation of the fly with minimum free parameters, our method uses a 3D model that includes two new aspects of wing deformation: A non-fixed wing hinge and a twisting wing surface. The method is demonstrated for free and perturbed flight. Our method does not use prior assumptions on the kinematics apart from the continuity of the wing pitch angle. Hence, this method can be readily adjusted for other insect species.

Distinct forms of resonant optimality within insect indirect flight motors.

Insect flight motors are extraordinary natural structures that operate efficiently at high frequencies. Structural resonance is thought to play a role in ensuring efficient motor operation, but the details of this role are elusive. While the efficiency benefits associated with resonance may be significant, a range of counterintuitive behaviours are observed. In particular, the relationship between insect wingbeat frequencies and thoracic natural frequencies is uncertain, with insects showing wingbeat frequency modulation over both short and long time scales. Here, we offer new explanations for this modulation. We show how, in linear and nonlinear models of an indirect flight motor, resonance is not a unitary state at a single frequency, but a complex cluster of distinct and mutually exclusive states, each representing a different form of resonant optimality. Additionally, by characterizing the relationship between resonance and the state of negative work absorption within the motor, we demonstrate how near-perfect resonant energetic optimality can be maintained over significant wingbeat frequency ranges. Our analysis leads to a new conceptual model of flight motor operation: one in which insects are not energetically restricted to a precise wingbeat frequency, but instead are robust to changes in thoracic and environmental properties-an illustration of the extraordinary robustness of these natural motors.

circMbl functions in cis and in trans to regulate gene expression and physiology in a tissue-specific fashion.

Muscleblind (mbl) is an essential muscle and neuronal splicing regulator. Mbl hosts multiple circular RNAs (circRNAs), including circMbl, which is conserved from flies to humans. Here, we show that mbl-derived circRNAs are key regulators of MBL by cis- and trans-acting mechanisms. By generating fly lines to specifically modulate the levels of all mbl RNA isoforms, including circMbl, we demonstrate that the two major mbl protein isoforms, MBL-O/P and MBL-C, buffer their own levels by producing different types of circRNA isoforms in the eye and fly brain, respectively. Moreover, we show that circMbl has unique functions in trans, as knockdown of circMbl results in specific morphological and physiological phenotypes. In addition, depletion of MBL-C or circMbl results in opposite behavioral phenotypes, showing that they also regulate each other in trans. Together, our results illuminate key aspects of mbl regulation and uncover cis and trans functions of circMbl in vivo.

Measuring pupil size and light response through closed eyelids.

Monitoring pupillary size and light-reactivity is a key component of the neurologic assessment in comatose patients after stroke or brain trauma. Currently, pupillary evaluation is performed manually at a frequency often too low to ensure timely alert for irreversible brain damage. We present a novel method for monitoring pupillary size and reactivity through closed eyelids. Our method is based on side illuminating in near-IR through the temple and imaging through the closed eyelid. Successfully tested in a clinical trial, this technology can be implemented as an automated device for continuous pupillary monitoring, which may save staff resources and provide earlier alert to potential brain damage in comatose patients.

Oocyte aging is controlled by mitogen-activated protein kinase signaling.

Oogenesis is one of the first processes to fail during aging. In women, most oocytes cannot successfully complete meiotic divisions already during the fourth decade of life. Studies of the nematode Caenorhabditis elegans have uncovered conserved genetic pathways that control lifespan, but our knowledge regarding reproductive aging in worms and humans is limited. Specifically, little is known about germline internal signals that dictate the oogonial biological clock. Here, we report a thorough characterization of the changes in the worm germline during aging. We found that shortly after ovulation halts, germline proliferation declines, while apoptosis continues, leading to a gradual reduction in germ cell numbers. In late aging stages, we observed that meiotic progression is disturbed and crossover designation and DNA double-strand break repair decrease. In addition, we detected a decline in the quality of mature oocytes during aging, as reflected by decreasing size and elongation of interhomolog distance, a phenotype also observed in human oocytes. Many of these altered processes were previously attributed to MAPK signaling variations in young worms. In support of this, we observed changes in activation dynamics of MPK-1 during aging. We therefore tested the hypothesis that MAPK controls oocyte quality in aged worms using both genetic and pharmacological tools. We found that in mutants with high levels of activated MPK-1, oocyte quality deteriorates more rapidly than in wild-type worms, whereas reduction of MPK-1 levels enhances quality. Thus, our data suggest that MAPK signaling controls germline aging and could be used to attenuate the rate of oogenesis quality decline.

Two-dimensional flow of driven particles: a microfluidic pathway to the non-equilibrium frontier.

We discuss the basic physics of the flow of micron-scale droplets in 2D geometry. Our focus is on the use of droplet ensembles to look into fundamental questions of non-equilibrium systems, such as the emergence of dynamic patterns and irreversibility. We review recent research in these directions, which demonstrate that 2D microfluidics is uniquely set to study complex out-of-equilibrium phenomena thanks to the simplicity of the underlying Stokes flow and the accessibility of lab-on-chip technology.

Walking like an ant: a quantitative and experimental approach to understanding locomotor mimicry in the jumping spider Myrmarachne formicaria.

Protective mimicry, in which a palatable species avoids predation by being mistaken for an unpalatable model, is a remarkable example of adaptive evolution. These complex interactions between mimics, models and predators can explain similarities between organisms beyond the often-mechanistic constraints typically invoked in studies of convergent evolution. However, quantitative studies of protective mimicry typically focus on static traits (e.g. colour and shape) rather than on dynamic traits like locomotion. Here, we use high-speed cameras and behavioural experiments to investigate the role of locomotor behaviour in mimicry by the ant-mimicking jumping spider Myrmarachne formicaria, comparing its movement to that of ants and non-mimicking spiders. Contrary to previous suggestions, we find mimics walk using all eight legs, raising their forelegs like ant antennae only when stationary. Mimics exhibited winding trajectories (typical wavelength = 5-10 body lengths), which resemble the winding patterns of ants specifically engaged in pheromone-trail following, although mimics walked on chemically inert surfaces. Mimics also make characteristically short (approx. 100 ms) pauses. Our analysis suggests that this makes mimics appear ant-like to observers with slow visual systems. Finally, behavioural experiments with predatory spiders yield results consistent with the protective mimicry hypothesis. These findings highlight the importance of dynamic behaviours and observer perception in mimicry.

Airborne Acoustic Perception by a Jumping Spider.

Jumping spiders (Salticidae) are famous for their visually driven behaviors [1]. Here, however, we present behavioral and neurophysiological evidence that these animals also perceive and respond to airborne acoustic stimuli, even when the distance between the animal and the sound source is relatively large (∼3 m) and with stimulus amplitudes at the position of the spider of ∼65 dB sound pressure level (SPL). Behavioral experiments with the jumping spider Phidippus audax reveal that these animals respond to low-frequency sounds (80 Hz; 65 dB SPL) by freezing-a common anti-predatory behavior characteristic of an acoustic startle response. Neurophysiological recordings from auditory-sensitive neural units in the brains of these jumping spiders showed responses to low-frequency tones (80 Hz at ∼65 dB SPL)-recordings that also represent the first record of acoustically responsive neural units in the jumping spider brain. Responses persisted even when the distances between spider and stimulus source exceeded 3 m and under anechoic conditions. Thus, these spiders appear able to detect airborne sound at distances in the acoustic far-field region, beyond the near-field range often thought to bound acoustic perception in arthropods that lack tympanic ears (e.g., spiders) [2]. Furthermore, direct mechanical stimulation of hairs on the patella of the foreleg was sufficient to generate responses in neural units that also responded to airborne acoustic stimuli-evidence that these hairs likely play a role in the detection of acoustic cues. We suggest that these auditory responses enable the detection of predators and facilitate an acoustic startle response. VIDEO ABSTRACT.

Wing-pitch modulation in maneuvering fruit flies is explained by an interplay between aerodynamics and a torsional spring.

While the wing kinematics of many flapping insects have been well characterized, understanding the underlying sensory, neural, and physiological mechanisms that determine these kinematics is still a challenge. Two main difficulties in understanding the physiological mechanisms arise from the complexity of the interaction between a flapping wing and its own unsteady flow, as well as the intricate mechanics of the insect wing hinge, which is among the most complicated joints in the animal kingdom. These difficulties call for the application of reduced-order approaches. Here this strategy is used to model the torques exerted by the wing hinge along the wing-pitch axis of maneuvering fruit flies as a damped torsional spring with elastic and damping coefficients as well as a rest angle. Furthermore, we model the air flows using simplified quasistatic aerodynamics. Our findings suggest that flies take advantage of the passive coupling between aerodynamics and the damped torsional spring to indirectly control their wing-pitch kinematics by modulating the spring parameters. The damped torsional-spring model explains the changes measured in wing-pitch kinematics during roll correction maneuvers through modulation of the spring damping and elastic coefficients. These results, in conjunction with the previous literature, indicate that flies can accurately control their wing-pitch kinematics on a sub-wing-beat time scale by modulating all three effective spring parameters on longer time scales.

Pitch perfect: how fruit flies control their body pitch angle.

Flapping insect flight is a complex and beautiful phenomenon that relies on fast, active control mechanisms to counter aerodynamic instability. To directly investigate how freely flying Drosophila melanogaster control their body pitch angle against such instability, we perturbed them using impulsive mechanical torques and filmed their corrective maneuvers with high-speed video. Combining experimental observations and numerical simulation, we found that flies correct for pitch deflections of up to 40 deg in 29±8 ms by bilaterally modulating their wings' front-most stroke angle in a manner well described by a linear proportional-integral (PI) controller. Flies initiate this corrective process only 10±2 ms after the perturbation onset, indicating that pitch stabilization involves a fast reflex response. Remarkably, flies can also correct for very large-amplitude pitch perturbations--greater than 150 deg--providing a regime in which to probe the limits of the linear-response framework. Together with previous studies regarding yaw and roll control, our results on pitch show that flies' stabilization of each of these body angles is consistent with PI control.

Controlling roll perturbations in fruit flies.

Owing to aerodynamic instabilities, stable flapping flight requires ever-present fast corrective actions. Here, we investigate how flies control perturbations along their body roll angle, which is unstable and their most sensitive degree of freedom. We glue a magnet to each fly and apply a short magnetic pulse that rolls it in mid-air. Fast video shows flies correct perturbations up to 100° within 30 ± 7 ms by applying a stroke-amplitude asymmetry that is well described by a linear proportional-integral controller. For more aggressive perturbations, we show evidence for nonlinear and hierarchical control mechanisms. Flies respond to roll perturbations within 5 ms, making this correction reflex one of the fastest in the animal kingdom.

Various p53 mutant proteins differently regulate the Ras circuit to induce a cancer-related gene signature.

Concomitant expression of mutant p53 and oncogenic Ras, leading to cellular transformation, is well documented. However, the mechanisms by which the various mutant p53 categories cooperate with Ras remain largely obscure. From this study we suggest that different mutant p53 categories cooperate with H-Ras in different ways to induce a unique expression pattern of a cancer-related gene signature (CGS). The DNA-contact p53 mutants (p53(R248Q) and p53(R273H)) exhibited the highest level of CGS expression by cooperating with NFκB. Furthermore, the Zn(+2) region conformational p53 mutants (p53(R175H) and p53(H179R)) induced the CGS by elevating H-Ras activity. This elevation in H-Ras activity stemmed from a perturbed function of the p53 transcription target gene, BTG2. By contrast, the L3 loop region conformational mutant (p53(G245S)) did not affect CGS expression. Our findings were further corroborated in human tumor-derived cell lines expressing Ras and the aforementioned mutated p53 proteins. These data might assist in future tailor-made therapy targeting the mutant p53-Ras axis in cancer.

Damped oscillations in the adaptive response of the iron homeostasis network of E. coli.

Living organisms often have to adapt to sudden environmental changes and reach homeostasis. To achieve adaptation, cells deploy motifs such as feedback in their genetic networks, endowing the cellular response with desirable properties. We studied the iron homeostasis network of E. coli, which employs feedback loops to regulate iron usage and uptake, while maintaining intracellular iron at non-toxic levels. Using fluorescence reporters for iron-dependent promoters in bulk and microfluidics-based, single-cell experiments, we show that E. coli cells exhibit damped oscillations in gene expression, following sudden reductions in external iron levels. The oscillations, lasting for several generations, are independent of position along the cell cycle. Experiments with mutants in network components demonstrate the involvement of iron uptake in the oscillations. Our findings suggest that the response is driven by intracellular iron oscillations large enough to induce nearly full network activation/deactivation. We propose a mathematical model based on a negative feedback loop closed by rapid iron uptake, and including iron usage and storage, which captures the main features of the observed behaviour. Taken together, our results shed light on the control of iron metabolism in bacteria and suggest that the oscillations represent a compromise between the requirements of stability and speed of response.

Burgers shock waves and sound in a 2D microfluidic droplets ensemble.

We investigate the collective motion of a two-dimensional disordered ensemble of droplets in a microfluidic channel far from equilibrium and at Reynolds number approximately 10(-4). The ensemble carries ultraslow shock waves and sound, propagating at approximately 100 microm s(-1) and superposed on diffusive droplets motion. These modes are induced by long-range hydrodynamic dipolar interactions between droplets, the result of the symmetry breaking flow. The modes obey the Burgers equation due to a local coupling between droplets velocity and number density. This stems from a singular effect of the channel sidewall boundaries upon summation of the hydrodynamic interaction in two dimensions.

Anomalous microfluidic phonons induced by the interplay of hydrodynamic screening and incompressibility.

We investigate the acoustic normal modes ("phonons") of a 1D microfluidic droplet crystal at the crossover between 2D flow and confined 1D plug flow. The unusual phonon spectra of the crystal, which arise from long-range hydrodynamic interactions, change anomalously under confinement. The boundaries induce weakening and screening of the interactions, but when approaching the 1D limit we measure a marked increase in the crystal sound velocity, a sign of interaction strengthening. This nonmonotonous behavior of the phonon spectra is explained theoretically by the interplay of screening and plug flow.

An autonomous system for identifying and governing a cell's state in yeast.

We present an approach for an autonomous system that detects a particular state of interest in a living cell and can govern cell fate accordingly. Cell states could be better identified by the expression pattern of several genes than of a single one. Therefore, autonomous identification can be achieved by a system that measures the expression of these several genes and integrates their activities into a single output. We have constructed a system that diagnoses a unique state in yeast, in which two independent pathways, methionine anabolism and galactose catabolism, are active. Our design is based on modifications of the yeast two-hybrid system. We show that cells could autonomously report on their state, identify the state of interest, and inhibit their growth accordingly. The system's sensitivity is adjustable to detect states with limited dynamic range of inputs. The system's output depends only on the activity of input pathways, not on their identity; hence it is straightforward to diagnose any pair of inputs. A simple model is presented that accounts for the data and provides predictive power. We propose that such systems could handle real-life states-of-interest such as identification of aberrant versus normal growth.