Internal Seminar Series

The OIST Internal Seminar Series is an informal meeting within the OIST community, where students, postdocs, research assistants and professors alike get to share a piece of their daily lives over free pizza/pastry. During a typical seminar, a speaker will introduce their unit or research field, discuss their latest results, or explain plans for future work.

One of the goals of the Internal Seminar Series is to foster an interdisciplinary, cooperative environment within OIST, where we can learn from each other and understand the motivations that drive research in fields we are not familiar with. This is based on the understanding that the activities at OIST are intrinsically interesting and worthwhile to learn about for curiosity's sake. Beyond this, there is the possibility that internal collaborations may spring up in unforeseen areas of research. We value the diversity of speakers: from students to professors, regardless of study field, age, race, and sex.

Seminars are organized every two weeks. Please see the "Calendar" or "Upcoming Events" links on the left for more information about upcoming seminars. Each seminar usually takes place in Lab 3 room C700 and starts at 16:00, after which there will be free pizza and beverages. The refreshment is financially supported by President's office, while they do not influence the topics, speakers, time, and venue of the seminars.

Note on alcohol:
From Fall 2024, we offer beer along the other refreshments. This is paid for privately by generous faculty and executives. Don't ruin it. Don't drink and drive. You can always call e.g. Ishikawa Taxi (098-972-5406) or Agachi Daiko (080-6485-9978).

Interested in speaking at an Internal Seminar? Fill out the form or contact internalseminar@oist.jp

2025 Winter

All seminars this term will be available both on-site and on Zoom.

February 4 (L5D23)

  • Ian (Cheng Yi) Hou, Organic and Carbon Nanomaterials Unit
Pentagon-Heptagon Pair in the Sea of Hexagons: from Azulene to the Dislocation Defects in Graphene
 
Pentagon, hexagon, and heptagon, these are several of the most fundamental geometric shapes, yet the geometrical arrangement of the π-electrons of sp2-carbon materials can have a dramatic influence on their chemical and physical properties. The ten-carbon isomers naphthalene and azulene, with skeletons of fused double-hexagon and pentagon-heptagon pair, respectively, is one of the most vivid examples. Going from the molecular world to the sea of hexagons in graphene, pentagon-heptagon pair is not an alien. Pentagon-heptagon pair is the nanostructure at dislocation defects in graphene, and it further accumulates and aligns at grain boundaries. In this talk, I will introduce our endeavors in dynamic azulene chemistry as well as attempts and plans in bottom-up construction of nanostructures related to dislocation defects in graphene.
 
  • Akira Kawano, Biological Physics Theory Unit

Decomposing Information Flow in Social Dynamics

Social behaviors include some of the most interesting interactions in living systems, yet their characterization is often qualitative. Here, we analyze social dynamics using information theory in the context of two male zebrafish engaged in a dominance contest (O’Shaughnessy et al, PRX Life 2024). Using 3D velocity data, we apply a partial information decomposition (PID) to identify modes of interactions in the fish pair. The analysis reveals patterns of redundant, synergistic, and unique information flow that unfold throughout the contest, offering a fresh perspective on dominance relationships. This work complements traditional game-theoretic approaches to social behavior with a quantitative, information-theoretic lens.

 

February 18 (L5D23)

  • Callum Hudson, Marine Climate Change Unit

Natural Analogues of Future Oceans: Using Volcanic CO2 Seeps to Study Sea Urchin and Ecosystem Responses to Climate Change

The oceanic uptake of anthropogenic carbon dioxide (CO₂) emissions is driving ocean acidification, a global environmental issue expected to negatively impact marine life. Aquarium-based studies have shown that organisms relying on carbonate (CO₃²⁻) to build shells and skeletons are particularly vulnerable to ocean acidification, which raises the energetic cost of calcification. However, a key question remains: can controlled experiments accurately reflect the complexity of natural systems, or might their findings differ under ecoloigclly realistic conditions? To better understand the effects of ocean acidification on organisms and ecosystems, researchers are increasingly moving beyond aquarium-based experiments on single species toward in-situ studies in naturally extreme environments, such as volcanic CO₂ seeps. These natural analogue sites have revealed that long-term exposure to acidification levels projected for the end of the century fundamentally alters marine communities. Complex, slow-growing habitat-forming species (e.g., corals and kelps) decline, while fast-growing, weedy species (e.g., algal turfs) proliferate, leading to simplified ecosystems with reduced biodiversity and ecological complexity. However, much of this research has been observational, leaving the ecological mechanisms driving these shifts unresolved. In this talk, I will discuss how naturally extreme environments can provide valuable insights into long-term community and species responses to climate change. I will then introduce my research on sea urchin responses to elevated CO₂ and the implications this will have for coastal ecosystems.

  • Ardak Kussainova, Mathematical and Theoretical Physics Unit

Dodecahedral spinel-type iron oxide NPs: low temperature synthesis and characterization

Recently arrived samples by JAXA Hayabusa2 mission from Ryugu asteroid of chondrite-type (ref 1) have renewed interest on well-known mineral, magnetite, especially in a view of processes that lead to the formation of space magnetite. The unique feature of the space magnetite, originated from the solar system, is its complex morphologies, which includes the shape of spheroids, plaquettes, euhedral crystals as icositetrahedrons. Such morphologies have previously also been seen in some meteorites as Orqueil and Ivuna (1). So far, it is important to emphasis that for magnetite to our best knowledge, such unique shape as 24-sided deltoidal icositetrahedron, was not found on Earth and could be considered as a feature of magnetite that originated in outer space (2). Taking into account the geological prerequisites as low temperature, we undertook a few syntheses by long heating the mixture of iron metal and deionized water in microtube at 65-75 C and in teflon autoclave at 180C. A combination of methods was carried out in this study. Based on x-ray diffraction (XRD) measurements and Rietveld refinement, scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS), high resolution TEM, Raman and SQUID magnetometry the formation of magnetite NPs in rhombic dodecahedral form was confirmed.
In this study we demonstrated that these experimental conditions favor magnetite formation and the direct water oxidation of metallic iron particles in a confined oxygen-free environment can provide a facile route to formation of stable spinel structure iron oxide NPs in a form of crystals with the 110 facets.

March 4 (L5D23)

  • Konstantinos Tsaridis, Computational Neuroethology Unit

Exploring hierarchical processing, polarization, and color perception in the cephalopod visual system


Cephalopods are marine invertebrates that have developed complex visual behaviors for hunting, communication, and most strikingly camouflaging. Unlike most vertebrates, cephalopods are seemingly color blind but have instead evolved to perceive the polarization of light. The neural mechanisms by which the cephalopod brain processes visual information to achieve these functions remain unclear. To study the functional organisation of their visual system, we developed a protocol for in vivo extracellular electrophysiology, using the bigfin reef squid, S. lessoniana, as a model.  During visual experiments that involved all three key properties of light—intensity, wavelength, and polarization—, we recorded neural population activity from throughout the Optic Lobe, the cephalopod visual processing center. We found that the cephalopod visual system, despite having independently evolved, shares many organization principles with the vertebrate visual system. Furthermore, our experiments provided insights into how polarization vision is utilized in underwater environments as an alternative to color vision.

 

March 18 (L5D23)

  • Daniel Muller-Komorowska, Neural Coding and Brain Computing Unit

Discovering Mechanisms and Biomarkers of Brain Disorders Using Simulation-Based Inference

Simulators are sophisticated computer programs that imitate the mechanisms of real-world systems and reproduce their output. Neuronal network simulators can incorporate mechanisms and dynamics of healthy and diseased brain areas. This makes them efficient tools for exploring disease causes and discovering compensatory mechanisms that restore healthy activity. However, modeling biological mechanisms is difficult because of degeneracy, the ability of biological systems to perform the same function in different ways. Because of degeneracy, even a simulator reproducing the system's output perfectly might be a poor representation of the biological system. Therefore, optimization techniques that find one optimal simulator are inadequate to discover biological mechanisms. Instead, we use simulation-based inference, a broad class of tools that estimate probability distributions of parameters given simulator output. Neural posterior estimation is a modern simulation-based inference method that requires fewer simulations than traditional methods, and it works well on both single-neuron and neuronal network simulators. Using neural posterior estimation, we found that the best possible compensatory mechanisms to maintain healthy neuronal activity depend on the underlying disease mechanisms. In the future, these disease mechanisms could be diagnosed in individual patients with the help of simulators. Our results show that spiking neuronal network simulators could provide the quantitative foundation to identify and treat specific pathophysiological network mechanisms with precise interventions.

 

April 1 (L5D23)

  • Simone Tandurella, Complex Fluids and Flows Unit

A multiscale perspective in particle-laden flows

Both in natural and artificial environments, flows can transport particles. These particles may be solid (river sediment) or liquid (rain, or droplets in the clouds), big (weather balloons) or small (volcanic ash), heavy (sand) or light (soap bubbles). The great diversity of possible flows can be in part reduced thanks to classical physics knowledge, but still remains vast and often difficult to probe. In this talk I aim to present the complexity of particle-laden flows, approaching the subject using a multiscale perspective. We will look at the phenomenology of single particles and larger assemblies, and finally at that of concentrated particle-laden flows. This will highlight the intertwined multiscale nature of particulate flows, but also some statistical ways how complex and often chaotic objects such as these are studied, described, and to some extent, understood. Throughout the explanation, I will include relevant results from my projects on the topic.

  • Payal Shah, Science and Technology Group
Conservation planning in an uncertain world: Lessons from biodiversity conservation in the Appalachian ecoregion
 
Climate change presents significant challenges to biodiversity conservation, making planning and resource allocation more complex. Traditional conservation approaches often overlook climate uncertainty, limiting their effectiveness in long-term biodiversity protection. We use a diversification-based approach to improve conservation strategies in the Appalachian ecoregion—one of North America’s most biodiverse landscapes. The region faces growing threats from habitat loss, climate change, and economic pressures on land use which introduces uncertainty into both ecological outcomes and conservation costs. We apply a portfolio optimization approach to design conservation policies that strategically allocate limited resources across species and locations while maintaining flexibility to adapt to climate and economic uncertainty. By evaluating multiple future scenarios, this framework identifies conservation investment strategies that balance trade-offs between biodiversity goals and climate-induced risks while allowing for dynamic resource reallocation over time.
 

April 15 (L5D23)

  • Aditya Singh, Nonlinear and Non-equilibrium Physics Unit

Scattering of standing waves by a draining vortex: Experiment

I will be discussing the experimental investigation into the interaction between a draining bathtub vortex and standing surface waves. Specifically, we examine how the presence of this vortex alters the structure of the standing waves, leading to the formation of distinct patterns that we refer to as "propellers" which radiate outward from the vortex core. The experimental setup involves generating a controlled draining vortex in a water tank and creating standing surface waves using transducers. By carefully tuning the parameters of the system, we measure the scattering effects and analyze the resulting wave patterns using high-speed imaging. Theoretical analysis complements these experiments, allowing us to model the interaction between the vortex and the waves, and to predict the conditions under which specific wavefront dislocations—and thus specific numbers of propeller blades—will occur. One of the key findings of our work is that the number of propeller blades—i.e., the arms or radiating spokes of the wavefront pattern—is governed by a dimensionless parameter, denoted as α. This parameter α contains information about the strength of the vortex and the wavelength of the surface waves. By varying α, we are able to control the number of propeller blades and observe different wavefront patterns.
 
  • Jonas Roenning, Nonlinear and Non-equilibrium Physics Unit
Emergent nodal lines in a classical analog of the Aharonov-Bohm effect: Theory
 
We present a theoretical study of the scattering of standing surface waves on a draining bathtub vortex. When the waves interact with the vortex, spinning propeller like lines are formed upon which the wave vanishes. These zero-lines run radially through the vortex and act as domain walls that separate regions where the waves have different phases. Their origin is similar to the wave front dislocations discussed in the fluid dynamic Aharonov-Bohm (AB) effect with traveling waves (Berry et al. European Journal of Physics, 1980) as an analog to the scattering of a beam of charged particles on a magnetic vector potential. In this classical analog, the height of the water-air interface corresponds to the real part of the quantum mechanical wave function, which is not a physical observable in the quantum case but can be observed in the classical setting. We consider the same analog quantum system using two oppositely directed beams to represent the standing wave. From the asymptotic wave function, one can predict the number of propellers (and their behaviour), which is determined by a dimensionless parameter α. These theoretical results have received both experimental and numerical confirmation. Like the wave front dislocation of the classic AB effect, these propellers are only observable in the classical case.
 
 

2024 Fall

September 10 (L4E48)

  • Dr. Alec LaGrow, Scientific Imaging Section
  • Dr. Eugene Khaskin, Science and Technology Group

September 24 (C700)

  • Dr. Melissa Matthews, Molecular Cryo-Electron Microscopy Unit

October 8 (C700)

  • David O'Connell, Quantum Gravity Unit

Space, Time and Sliding Doors

In this talk I will cover a few elements of my PhD project, which roughly deals with topology change in a physical context. Starting with an observation surrounding Hollywood's dearest Gwyneth Paltrow, we will slowly walk through the relevant structures required to phrase my PhD research goal. Along the way we will discuss some aspects of topology, geometry and relativity theory, hopefully with a bit of something for everyone.

  • Giulio Foggi Rota, Complex Fluids and Flows Unit

Turbulent flows over flexible canopies

Canopy flows occur when numerous slender elements protruding from a supporting surface interact with a moving fluid. These flows are often found in nature and are particularly interesting from geophysical and biological perspectives. In this work, we begin by examining the dynamics of a flexible fibre clamped in turbulent flow and then extend this understanding to the collective behaviour of flexible canopies. This development provides fundamental physical insights that pave the way for investigating more complex and realistic natural scenarios, such as the swaying of crops in the wind or the agitation of kelp forests by marine currents.

October 22 (C700)

  • Ryo Nakatani, Computational Neuroscience Unit

Astrocytes actively alter electrical membrane potential; How and why

The electrical properties of cellular membranes are the basis for information transfer within the brain. Typically, these processes are done with the intercommunication of neurons via dynamic changes of the electrical membrane potential (hyper-polarization/depolarization). However, recent studies have also observed that astrocytes, non-neuronal supportive cells, also have depolarization, specifically in the periphery where they are in close contact with neurons. Utilizing a computational whole-cell astrocyte model, we explore previous hypotheses on the mechanism of astrocytic depolarization as well as provide a new prediction into how these mechanisms work, and why they are relevant.

  • Hirokazu Maruoka, Nonlinear and Non-equilibrium Physics Unit

Data-driven discovery of self-similarity using neural networks

Physics is the fundamentally a data-driven enterprise based on the experimental data and simulation data. In this seminar, I will talk about our powerful data-driven approach that combines self-similarity and neural networks. By introducing similarity parameters, which are power-law monomials composed of physical parameters, all the data points converge to a single line, which is known as data collapse. This means that all the self-similarity of data points is fully exploited, providing significant insights into the physics. In order to obtain these data collapse relations, the only unknown parameters are their power exponents of the similarity parameters. The general conventional method to obtain power exponents has been dimensional analysis though it is limited to only special cases. Based on this insight of self-similarity, we have succesfully invented a systematic approach to get data collapse using neural network technique. Through this seminar, I would like to demonstrate : what is data-collapse? what is self-similarity? and what is the strategy to get data-collapse from arbitrary data?

November 5 (C700)

  • Dr. Jamila Rodrigues and Dr. Yi Huang, Complexity Science and Evolution Unit

Okinawa coastal fisheries: exploring ideas of climate change and the future of fishing industry

Japan's coastline is characterised by the presence of fishing communities that play a pivotal role in the provision of seafood. These community-based fisheries provide employment for local families while simultaneously embodying cultural heritage. Just like other fishing communities worldwide, Okinawa fishers are confronted with challenges such as the impact of climatic stressors that affect marine ecosystems and the fishing industry. In this seminar, we introduce our project focusing on Okinawa fishers. We have developed a transdisciplinary study in collaboration with scholars, local fishers, and the Okinawa Prefecture Fisheries Cooperative Federation. The objective of our study is to analyse fishers’ perceptions of local climate change, their concerns about the future of the fisheries industry, and their notions of well-being. By incorporating a transdisciplinary approach, our study aims to provide a broader understanding of how local climate change is perceived by local fishers and its impacts on Okinawa fishing communities. We will introduce our research methods and some preliminary findings. We are open to discuss with the audience on how to develop strategies for researchers who are interested in engaging with coastal communities and beyond.

  • Zohreh Shahrabifarahani, Light-Matter Interactions for Quantum Technologies Unit

Two photon generation using optical nanofiber-trapped cold atoms useful for “Heralded single photon” scheme 

We aim to experimentally realize the nonlinear effect of Four-Wave Mixing (FWM) in an atomic system cooled and trapped near an optical nanofiber (ONF) waveguide. This setup enables the generation of pairs of correlated photons, which are beneficial for heralded single-photon source schemes. The optical nanofiber facilitates this process by allowing for significantly lower power requirements and enabling efficient coupling of the generated photons back into the ONF for transmission to any desired destination. The produced atomic-based single photons will exhibit a narrowband spectrum due to the cancellation of the Doppler effect in cold atoms, and compatible with other atomic-based quantum systems, including atomic quantum memories.

November 19 (C700)

  • Lenard Lajos Szantho, Model-Based Evolutionary Genomics Unit

Towards better phylogenetic models of evolution by accounting for compositionally constrained sites

Accurate phylogenies are fundamental to our understanding of the pattern and process of evolution. Yet, phylogenies at deep evolutionary timescales, with correspondingly long branches, have been fraught with controversy resulting from conflicting estimates from various phylogenetic models. Analyses of empirical datasets have demonstrated that inadequate modeling of across-site compositional heterogeneity, which is the result of biochemical constraints that lead to varying patterns of accepted amino acids along sequences, can lead to erroneous topologies, that are strongly supported, e.g. observing long-branch attraction (LBA) artifact. Unfortunately, models that adequately account for across-site compositional heterogeneity remain computationally challenging or intractable for an increasing fraction of contemporary datasets.

In my talk I will introduce "compositional constraint analysis" a method to investigate the effect of site-specific constraints on amino acid composition on phylogenetic inference, and "CAT-PMSF", a phylogenetic pipeline to ameliorate the effect of LBA artifacts by accounting for across-site compositional heterogeneity while still running in a feasible amount of time as presented on several empirical datasets including Microsporidia, Nematoda, and Platyhelminthes.

  • Emily Satyadhi, Molecular Neuroscience Unit

Impairment in axonal translation caused by reduction in ATP level during aging in sensory neurons

Axonal translation is an important mechanism which plays a role in maintaining axonal morphology as well as mediating axonal injury and recovery. During axonal translation, mitochondria is thought to be stored along the axon to provide energy required for translation. The decline in mitochondria activity is one of the hallmarks of aging, however, it is still unclear whether the decline in mitochondria activity can be linked to the level of axonal translation in aging neurons. In here, we utilize the microfluidic device to separate the cell body and the axon of the DRG neurons. Using live imaging, we found that there is a significant decrease in the level of axonal translation level as well as the number of translational hotspots in aging neurons. We also showed that supplementing the culture with NMN, which can increase mitochondria activity, was able to increase the level of axonal translation of the aging neurons similar to the level observed in young neurons.

December 3 (C700)

  • Emma Gairin, Marine Eco-Evo-Devo Unit
Decoding the relationship between fish gene expression and the environment
 
A large fraction of the worldwide human population lives within a few kilometers of the oceans, causing unprecedented degradation to coastal habitats. This degradation is often monitored with water parameters or biological community surveys, but these are not directly informative of the compounded impact of stressors on organisms. Transcriptomics may offer an exhaustive method to assess the effects of human activities on organisms by assessing the expression level of all genes, revealing sublethal effects and acclimation to local conditions. However, this is hampered by the versatility of gene expression, which is often considered to hide environmental effects. Here, we obtained the transcriptomes of adult livers and whole-body juveniles of the blue damselfish Chrysiptera cyanea in eighteen coastal locations, from urban bays to isolated reefs, in Okinawa, Japan. We confronted the field data with laboratory-generated transcriptomes of fish exposed to a single stressor. Using pathway reconstruction, we show that the transcriptomic landscapes were influenced by urbanization level and detected a specific signature for marker genes associated with chemical pollution exposure. By comparing the signature of fed and fasted fish in the laboratory with changes in energy metabolism in the various field locations, we were able to retrieve the nutritional status of each fish and identify that urban zones may provide ‘fast food’, with large quantities of low-quality nutrients. Taken together, our results demonstrate that environmental transcriptomics coupled with pathway reconstruction and lab experiments can provide unsuspected information for ecosystem assessment and reveal the profound and long-lasting impact of urbanization on coastal fishes.
 

December 17 (C700)

  • Dr. Florian Pflug, Biological Complexity Unit
Individual growth vs. population growth in E. coli, a problem of optimal resource allocation 
 
In a growing population of bacteria we observe two separate but interlinked exponential growth processes. The first, well-known process is the exponential increase in number of cells, driven by each cell's cycle of genome replication, growth and division. The second is the exponential growth of cell sizes, driven by the self-reproduction of ribosomes. To maximise their fitness, bacteria must find an optimal balance between these two processes. Finding this balance can be framed as the problem of optimal allocation of resources between DNA replication (which is a prerequisite for division) and ribosome production (which drives cellular growth). We present a mathematical model for this problem of resource allocation, and show that can predict the change in growth rate in response to a change resource allocation. To validate this model, we present an experimental method that allows us to precisely measure the timing of replication in E. coli which serves as a stand-in for the amount of resources allocated to this process. We apply this method to different E. coli mutants with perturbed allocation of resources, and show that our model predictions agree reasonably well with experimental observations.
 
  • Dr. Samy Lakhal, Nonlinear and Non-equilibrium Physics Unit

On random fields and the laws of nature

The shapes, forms and textures surrounding us are rich in details. In many physical setups, that richness actually emerges from the taming of noise and fluctuations by the physical laws of nature. By mimicking these laws, scientists, graphic designers and other artists can reproduce complexity at small costs. In this talk, I will introduce some basics of random modeling and their applications to texture synthesis. I will then show direct examples and applications in physics and visual design.
 
 
Current organizers:
  • Yasha Neiman (yaakov.neiman@oist.jp) 
  • Anouk Beraud (a.beraud@oist.jp)
  • Monika Eggenberger (monika.eggenberger2@oist.jp)
  • Aleksandra Gavrilova (aleksandra.gavrilova@oist.jp)
  • Aditya Singh (aditya.singh@oist.jp)

Reach all current organizers at: internalseminar@oist.jp

Special thanks to Akiko Ishimine (President Office) for managing the budget and payment for internal seminars, Tomoko Yoshino (Mitarai Unit) for dealing with paper works for the payment, and Akiko Mizukoshi (University Community Services) for communicating with the cafe to provide refreshment.

Former organizers:

  • Jonathan Ward
  • Yunhui Zheng
  • Amy Shen
  • Eliot Fried
  • Stefan Pommer
  • Carolina Diaz
  • Tina Mularski
  • Oleg Simakov
  • Alexander Badrutdinov 
  • Omar Jáidar 
  • Marco Tsui 
  • Eiji Uchibe 
  • Hiroshi Izumi
  • Jérémie Gillet
  • Mary Ann Price
  • Shao Fang Wang
  • Romanas Chaleckis
  • Bianca Sieveritz
  • Lauren Dembeck
  • Robert Campbell
  • Andreas Thomasen
  • Jason Ball 
  • Maéva Techer 
  • Maggi Mars Brisbin
  • Afshan Jamshaid 
  • Fabien Benureau
  • Nadine Wirkuttis
  • Kamila Mustafina
  • Yuka Suzuki
  • Kun-Lung Li
  • Jigyasa Arora
  • Lewis Ruks
  • Aleksandra Bliznina
  • Ainash Garifullina
  • Thato Mokhothu
  • Mohamed Boubakour
  • Lakshmipriya Swaminathan
  • Munissa Sadykova
  • Charles Whitaker
  • Akira Kawano
  • Marylka Yoe Uusisaari