2018 Simons Society of Fellows Retreat

It is not strictly within the duties of junior fellows studying the intricacies of Riemannian manifolds to inspire CRISPR-expert fellows to go in new research directions, or vice versa.Nonetheless, the atmosphere of the Simons Society of Fellows (SSF) retreats certainly nurtures such contributions and smooths scientific communication across disciplines.

Starting in 2017, the SSF has gotten together for a weekend at an idyllic location to share the latest research advances of both junior and senior fellows. It also gives fellows the opportunity to escape the pace of the city and polish junior fellows’ interpersonal skills by interacting with accomplished scientists. The outcome of that rather unique interaction is a valuable, restful and liberating experience.

Bluffton, South Carolina, served as the perfect venue for the 2018 SSF retreat. The southern charm and temperate late-February weather set the stage for 17 talks, four of them given by senior fellows. Below is a summary of those talks:

Arkarup Banerjee opened the meeting by sharing evidence for the biological basis of behavior. He described how the complex vocalization of Scotinomys teguina (Central American cloud forest rodents) offer exceptional clues to how motor outputs are connected to sensory stimulations. Arkarup’s efforts at the NYU Langone Medical Center could certainly uncover new aspects of social and vocal interactions.

Shana Caro followed with an interesting evolutionary study of honest signals. She illustrated the intricate interplay between siblings, food abundance and adequate bird parental care. For instance, offspring begging could easily fail as a survival strategy if family resources are scarce, hence emphasizing the importance of matching parent-offspring evolutionary interests. Shana’s current research interests focus on cooperative breeding.

Sylvain Carpentier described a few essential mathematical elements of integrable systems. He shared examples of particular interest in physics, where stable wave forms describe complex responses. Some systems are classical, like mechanical systems with finite dimensions; however, extending dimensionality to infinity offers vast opportunities for mathematical explorations. Sylvain is actively engaged in exploring such ‘mysteries.’

Eric Castillo gained fellows’ attention with his evolutionary posture toward modern spinal pathologies. He shared extensive experimental evidence for variant lumbar curvatures connected to bipedalism. Some other fascinating examples connecting barefoot runners and development provided more elements that contribute to a better understanding of who we are. Eric’s contributions could also expand our approaches to efficiently treating diseases.

Sara Fenstermacher shared her contributions to neural control of motor behavior. She illustrated strategies to map serotonergic neurons and their function in the spinal cord. Sara’s contributions are helping reveal the specific neural organization and roles that make us efficiently coordinate movement and reflex response.

Ruth Angus took the fellows through a voyage across space and time. She described how much of what we know about planetary systems is highly connected to our understanding of their host stars. By knowing some properties of planets such us radii, temperatures and orbital periods, it is possible to reconstruct an image of exoplanet evolution and perhaps predict their fate. Ruth will continue to lead such efforts examining large data sets.

Takashi Onikubo offered an engaging lecture on the mechanistic basis of gene transcription distal enhancers. In particular, he shared experiments that help understand the fundamental role of cohesin during transcription. Cohesin-DNA interactions entail dynamic conformational changes that seem to explore multiple states before finding an optimal configuration (linking two distal DNA sequences). Takashi continues to explore the transcription mechanism in detail.

Eliran Subag introduced mathematical models dealing with random functions in high-dimensional space. The complexity of the mathematical treatment is helping understand physical systems with highly irregular properties such as disordered magnets (spin glasses). Eliran’s work can be extended to study dynamic energy landscapes and the presence of metastability regions.

Yi Sun illustrated applications of random covariance matrices. One of the most well-known applications is principal component analysis (PCA), which is being actively used in experimental analysis to decorrelate data. Yi’s mathematical contributions could extend our capabilities in handling large, high-dimensional data sets.

Carlotta Ronda presented details behind MAGIC (metagenomic alteration of gut microbiome by in situ conjugation). Carlotta’s technique reveals and controls the complex dynamics of microbial communities with the capability of in situ modification of gut microbiomes. Opportunities to reengineer host physiological response could be within reach by efficiently accessing gut microbiome balance.

Bianca Jones Marlin offered a talk on how olfactory fear conditioning in mice could be passed along to their offspring — transgenerational epigenetic inheritance. She continues to identify biological mechanisms that could orchestrate the efficient transmission of epigenetic information.

Keith Hawkins shared his work as a galactic archeologist. Given the vast distances in our universe, a simple star survey turns the objects of study into massive fossils. Keith contributes to understanding the evolution of galaxies in our universe with the help of large spectroscopic surveys (Gaia). He hopes to provide us with an integral picture of galactic chemical distributions.

Michael Waskom closed the junior fellows’ talks by illustrating advances to deciphering the neural basis of decision making. He explained how accumulated evidence leads to commitment or choice. Michael records neural activity with functional magnetic resonance imaging to better understand the complexity of decisions at multiple time scales.

Senior fellows took the stage to share scientific contributions that have matured over time, turning into field paradigms. Boris Altshuler shared some principles of physical integrable systems and special cases of localization. Margaret Wright offered two snapshots: the first regarding simplex models, and the second based on real-world applications for the YRMH-IGYT model. Carol Mason illustrated our current understanding of development in higher vertebrates’ visual pathways and regenerative potentials for proper binocular vision. Moses Chao described evidence for exercise-induced metabolic changes connected to gene expression of a healthy brain.

 

Contributed by Jairo Diaz

Comments and edits by Tobias Bartsch and Tony Movshon

Arkarup Banerjee
NYU Langone Medical Center

Cortical Control of Vocal Communication in a Neotropical Singing Mice

The ability to act upon sensory information to generate a desired motor output is a fundamental component of animal behavior. During conversation, for example, we listen to the words of another person, interpret them and modify our speech appropriately. How the brain enables such flexible sensorimotor transformations is a central question in neuroscience. Carefully designed behavioral tasks with precise experimental control over both the stimuli as well as motor responses have advanced our understanding of neural mechanisms that support sensorimotor transformations. However, we know little about the neural mechanisms that govern more flexible and natural behaviors such as vocal communication between different members of a given species.

To address this issue, Banerjee and his colleagues recently started to investigate neural mechanisms of vocal communication in Alston’s singing mouse (Scotinomys teguina) – a highly vocal neotropical rodent native to the cloud forests of Central America. S. teguina produce a stereotyped series of vocalizations (in the human audible range) that are often performed in concert with the songs of other individuals of the same species. This vocal coordination has similarities to “turn taking” in human speech as well as vocal behaviors observed in nonhuman primates. The goal of Banerjee’s project is to investigate the neural dynamics that integrate the sensory (i.e., song of vocal partner) and motor (i.e., vocal response) aspects of vocal interactions, thereby gaining insights into ethologically relevant sensorimotor transformations. In his talk, he will describe the singing behavior of these rodents and discuss a series of experiments that were performed to localize a motor cortical area that influences vocal production and controls vocal coordination in this species.

 

Shana Caro
Columbia University

Neglectful Parents and Dishonest Offspring

Signals transmit information, and they have evolved across the tree of life in forms as diverse as singing in humpback whales, colorful skin in poison dart frogs, and quorum-sensing in bacteria. In order for any signal to evolve and be maintained, both signalers and receivers must gain an evolutionary benefit from communicating. When the evolutionary interests of signalers and receivers are not perfectly aligned, communication can be disrupted by the incentive to signal dishonestly or to ignore signals. Shana Caro investigates how signaling systems are influenced by the degree of conflict amongst communicators, and whether different kinds of stressors have different effects on communication. Caro uses parent-offspring communication (begging) in birds as a model. Offspring birds have evolved elaborate begging signals to solicit their parents for food, but families face an evolutionary conflict of interest over how parents should divide food among their chicks. Caro investigates how environmental variability, environmental quality, and life history traits influence the severity of within-family conflict, and how that conflict then shapes communication over evolutionary and ecological time.

 

Sylvain Carpentier
Columbia University

Why Should One Care about Integrable Systems?

This talk is an invitation to the theory of integrable systems, which are differential equations possessing a large group of symmetries. We will see how this small class of equations is ubiquitous in physics and how its study relates to various fields of mathematics, from algebraic geometry to pure analysis. Carpentier will explain his contribution to this theory, from the point of view of differential algebra.

 

Eric Castillo
Hunter College

Bipedalism and Back Pain: Insights into the Evolution and Function of the Human Lumbar Spine

From why we spike a fever when sick to what may have caused the rise in modern obesity rates, evolutionary theory can provide fresh perspectives on health and disease. In this talk, Eric Castillo will discuss how his research on the evolution of human lumbar lordosis (curvature of the lower spine) may shed new light on the prevalence of lower back pain today.

Lordosis was a key postural adaptation that helps humans balance and stabilize the mass of the upper body over two legs. But lordosis in modern populations and estimates from the fossil record suggest there has been substantial variation in the degree of lordosis through time. Given the evolutionary and clinical importance of lordosis, why is it so variable? Integrating lab experiments, fossil analyses, and fieldwork, this talk will explore the biomechanical consequences of postural variability in modern humans and the evolutionary pressures shaping the spines of early hominins. Finally,Castillo will discuss the implications of this research for understanding the etiology of spinal pathologies including back pain as well as new directions for research in evolutionary medicine.

 

Sara Fenstermacher
Columbia University

Neural Control of Movement: Serotonin in the Spinal Cord

The nervous system is capable of generating an incredibly vast repertoire of motor behavior. Every action-from holding a pencil, to speaking, to running up a set of stairs-requires the precise and coordinated activation of specific muscles, which is orchestrated by neural circuits within the brain and spinal cord. Serotonin, a chemical messenger synthesized in the brain, is released throughout the brain and spinal cord where it acts upon neurons and adjusts the output of neural circuits. While it is known that the motor system is influenced by serotonin, the organization and function of the serotonergic system with respect to motor circuits is not well understood. This work will examine the role of serotonergic modulation in the control of movement, with particular focus on spinal motor circuits. First, using viral and genetic strategies in the mouse, Sara Fenstermacher will identify the specific serotonergic neurons that participate in motor control. Second, she will manipulate components of the serotonergic system in the awake, behaving animal to investigate the function of serotonin in the spinal cord. Together, these studies will provide novel insight to the cellular mechanisms by which serotonergic pathways contribute to the breadth and flexibility of motor behavior.

 

Ruth Angus
Columbia University

Planetary Systems Across Space and Time

Thousands of planets beyond our own solar system have been discovered in the last two decades and it is now clear that exoplanets are abundant in our neighborhood of the Milky Way. Despite this large sample, characterizing trends in the planet population is challenging. This is partly because the properties of their host stars are poorly constrained. In particular, the way planetary systems evolve over time is still unknown because stellar ages are notoriously difficult to measure. Did certain epochs of the universe’s history produce more exoplanets than others? Is there a “sweet spot” in time and space for life to develop? These questions remain unanswered despite the huge, information-rich data sets available.

Ruth Angus’s research focuses on improving our understanding of stellar evolution, developing new age-dating methods for stars and exploring the exoplanet population by applying statistical techniques to large data sets. With a number of large-scale astronomical surveys scheduled to begin in the next few years, the ability to measure precise stellar ages will expose the evolving nature of planetary systems and reveal clues to the origins of life in our galaxy.

 

Takashi Onikubo
Rockefeller University

Role of Cohesin in Distal Enhancer Promoter Interactions and Transcription

Transcription is a cellular process in which the genetic information stored in DNA is copied onto RNA, which in turn serves as a blueprint for the synthesis of corresponding proteins. As all cells within an organism, with few exceptions, share the identical genome (the collection of all genetic information in DNA), cell identity arises from the selective transcription of a repertoire of genes unique to a given cell type. Transcriptional regulation is thus a process fundamental to life of an organism and to its constituent cells. Transcription is initiated by a family of proteins known as transcription factors or transcription activators, which binds to a specific DNA sequence within the genome (enhancers) and assembles/activates RNA polymerase complex (Pre-Initiation Complex) at the start site of transcription (promoters). However, these enhancers are often found at sites far distal to the promoters, reaching hundreds of kilobases (base: a single molecule of DNA) for some enhancers. Thus, physical linking of distal enhancers to promoters is a critical process to initiate and maintain proper target gene transcription for a given activator.

Takashi Onikubo’s research focuses on the mechanistic basis of this process with an emphasis on cohesin, a ring-shaped protein complex implicated in 3D organization of the genome, and focusing on how cohesin links two distally positioned stretches of DNA, enhancers and promoters, in the context of transcription. Here using DNA tethered to beads we show that cohesin complexed with its loader NIPBL interacts with DNA weakly and a second free stretch of DNA greatly increases its loading onto the tethered DNA. Our result suggests that the interaction with two stretches of DNA closes the ring and thereby traps two strands of DNA in a single ring of cohesin.

 

Eliran Subag
New York University

Random Functions on High-Dimensional Spheres

View Slides (PDF)

A function on a high-dimensional space, when chosen at random, is typically very complex. In this talk Eliran Subag will discuss such functions that arise in the study of spherical spin glass models – i.e., mathematical models for disordered magnetic systems. The main focus will be on geometric properties such as critical points and the structure of level sets, and how those relate to objects of interest from the physical perspective.

 

Margaret Wright
New York University

Two Snapshots from Numerical Optimization

View Slides for Snapshot 1 (PDF)

View Slides for Snapshot 2 (PDF)

Research in numerical optimization includes a variety of modes: developing new algorithms; analyzing algorithms in an attempt to understand convergence as well as their likely “in the large” behavior; creating mathematical models of real-world problems; and writing code that solves interesting problems. Margaret Wright will talk about two recent activities along some of these lines.

 

Yi Sun
Columbia University

A Probabilistic View on Random Covariance Matrices

Eigenvalues of random matrices were first studied by Wishart in the early 20th century for covariance estimation in mathematical statistics. Yi Sun will survey the study of these matrices from this and other mathematical perspectives and discuss some of his recent work studying such matrices using tools from special functions and free probability.

 

Carlotta Ronda
Columbia University

Metagenomic Engineering of the Mammalian Gut Microbiome in Situ

View Slides (PDF)

Current technology limitations in genetic tools for non-standard model organisms prevent a comprehensive understanding of microbial community dynamics since many microbes have yet to be (or cannot be) cultivated in the laboratory. Therefore, there is a need to develop new genetic techniques to engineer undomesticated microbes in order to expand our limited repertoire of microbial chasses that can be used to introduce heterologous relevant functions into cells that live in complex and challenging environments. The inability to genetically alter a bacterium greatly limits our basic understanding of the organism and its biotechnological potential. To overcome these challenges, we devised an approach, Metagenomic Alteration of Gut microbiome by In situ Conjugation (MAGIC), to genetically modify gut microbiota in its native habitat by engineering the mobilome – the repertoire of mobile genetic elements that permeate the gut microbiome. Implementing MAGIC directly on a complex microbiome in its natural habitat enables metagenomic infiltration of genetic payloads in microbes otherwise recalcitrant to laboratory cultivation or isolation of modifiable microbial chasses for synthetic applications. We introduced recombinant DNA in bacteria from the mammalian gut microbiome, identified and isolated recombinant carriers that were amenable to genetic manipulation, and then redeployed the recombinant bacteria back into the mammalian host as host-optimized engineerable probiotics. Furthermore, we demonstrated in situ gene transfer in a live animal to deliver new genetic traits directly into its established gut microbiota. This in situ genome engineering platform enables accelerated development of new microbial chasses for synthetic biology and the introduction of novel capabilities into established microbial communities with minimal disruption to their native milieu. In conclusion MAGIC opens the possibility of using a less invasive and more precise genetic approach to study complex dynamics in microbial communities and host- microbiome interactions.

This is joint work with Sway P. Chen, Vitor Cabral, Stephanie Yaung, Harris H. Wang. Department of Systems Biology, Columbia University.

 

Bianca Jones Marlin
Columbia University

Transgenerational Epigenetic Inheritance of Traumatic Memory

The past of our ancestors shapes our present. Whether in history, or in genetics, their experiences live on through us. Bianca Jones Marlin investigates the genetic, epigenetic, and transgenerational consequences of parental trauma on future generations. Specifically, she examines the mechanisms by which experience-dependent gene expression leads to changes and neural structure and behavior, and how these changes are transmitted across generations. Bianca Jones Marlin uses an olfactory fear-conditioning paradigm to address when, where, and how olfactory fear learning is coded in early olfactory processing. She hypothesizes that olfactory fear conditioning increases the salience of the conditioned odor through the proliferation of an undefined stem cell population in the main olfactory epithelium leading to neuroanatomical, physiological and behavioral changes. Jones Marlin posits olfactory receptor neurons responding to behaviorally salient cues are selected for by local non-coding RNA signals activated downstream of neuronal activity. She plans to examine the biological mechanisms by which odor information is transgenerationally inherited. She hypothesizes that transmission of epigenetic information from parent to sperm will recapitulate transmission from sperm to offspring. Bianca Jones Marlin’s research has shown that experience dependent changes in the main olfactory epithelial are a dynamic phenomenon, demonstrating that neural networks respond to salient contexts in a profile-specific manner.

 

Keith Hawkins
Columbia University

Galactic Archaeology: A Story of Our Milky Way Galaxy

One of the key objectives of modern astrophysics is to understand the formation and evolution galaxies. In this regard, the Milky Way is a fantastic testing ground for our theories of galaxy formation. However, dissecting the assembly history of the galaxy requires a detailed mapping of the structural, dynamical chemical, and age distributions of its stellar populations. Recently, we have entered an era of large spectroscopic and astrometric surveys, which has begun to pave the way for the exciting advancements in this field. Combining data from the many multi-object spectroscopic surveys, which are already underway, and the rich dataset from Gaia will undoubtedly be the way forward in order to disentangle the full chemo-dynamical history of our galaxy.

In this talk, Keith Hawkins will discuss his current work in galactic archaeology and how large spectroscopic surveys have been used to dissect the structure of our galaxy. He will also explore the future of galactic archaeology through chemical cartography.

 

Michael Waskom
New York University

Decision-Making through Evidence Accumulation at Multiple Timescales and Levels of Abstraction

Evidence accumulation models of decision-making propose that decisions are formed through a process where multiple samples of information (evidence) are temporally integrated (accumulation), producing a quantity that can be compared against a criterion to determine choice. This framework has several virtues: it is rooted in normative decision theory, it can account for a wide range of behavioral effects, and it corresponds to changes in neural activity during decision formation. It has therefore been proclaimed as a general mechanism that might underlie the complex decisions humans face in daily life. Yet scaling laboratory tasks to the duration and complexity of these naturalistic decisions poses many challenges for evidence accumulation. For example, naturalistic decisions often unfold over long and unpredictable timescales. Do these exceed the limits of biophysical mechanisms that integrate evidence in simple tasks? Naturalistic environments also require second-order decisions not just about the evidence itself but also about how best to use it. Can accumulation models be extended to explain these more abstract decisions? To address these questions, we are developing novel approaches for studying the computational and neural basis of decisions that vary in their timescale and complexity.

 

Carol Mason
Columbia University

Wiring the Embryonic Eye to the Brain for Binocular Vision

In higher vertebrates, proper binocular vision depends on the extension of retinal ganglion cell (RGC) axons at the optic chiasm to the same (ipsilateral) and opposite (contralateral) side of the brain. We have defined the molecular factors that direct the growth and guidance of RGCs during the development of this circuit and the genes that specify precursor cells to become ipsi- or contralaterally-projecting RGCs. We currently seek to uncover how time and place of retinal progenitor cell birth and differentiation are related to cell fate, including axonal projection. To further understand these developmental steps, we study the albino visual system in which the lack of pigment in the cell layer surrounding the eye is linked to a reduction of the ipsilateral RGC retinal projection; how this perturbation occurs is a long-standing conundrum. Carol Mason will briefly touch on the relevance of our work for driving stem cells into RGCs and directing axon regeneration to repair injured and degenerating visual pathways.

 

Moses Chao
New York University

How Does Exercise Help the Brain?

Exercise is not only beneficial for physical well-being, but also confers positive effects upon mental health and cognition. The molecular mechanisms that bring about these positive effects are not fully understood. We have discovered a link between metabolic changes resulting from exercise and increases in gene expression in the brain, by focusing upon the production of growth factors. A mechanistic understanding of physical exercise is relevant to neuropsychiatric problems, such as anxiety, depression and Alzheimer’s and Huntington’s disorders.