Selected Fellows Call 2015

Below you will find an overview of the proposals of the successful candidates of the 2015 call by alphabetical order.

To explore the scientific outputs of our fellows, please search them on Infoscience

Unravelling the molecular basis of local sequence-independent, regulatory variation in humans

Prof. Deplancke – Group UPDEPLA

See publications

©Masha Litovchenko

Explaining the genetic basis of complex traits or disease phenotypes is an outstanding challenge for biomedical research. Genome wide association studies showed that most common disease-associated genetic variants were found outside of gene-coding sequences, i.e. in the regulatory parts of the genome. This finding highlight the importance of elucidating how non-coding genetic variants mediate phenotypic variation.

A recent study performed in my host lab has revealed that regulatory variants can impact a coordinated change in transcription factors (TF) binding and chromatin modifications over regions spanning >100 kb, termed variable chromatin modules (VCMs). This pioneering concept, emphasizing a fine-grained, modular organization of the genome, may greatly enhance our understanding of how regulatory variation shapes complex traits.

The goal of my project is to study the universal and dynamic nature of VCMs. Using the state-of-art genome editing tools we will address the mechanisms of VCM activity and provide experimental evidence that a single TF-DNA interaction may impact the activity of an extended genomic region. We will also explore the plasticity of VCMs in the context of an evolving transcriptional program during human adipogenesis. Our results have great potential to contribute to the molecular interpretability of non-coding, genome-wide associated variants, which so far remain largely uncharacterized.

(Start date of fellowship: 1 March 2016)

©Daniel Alpern
Artistic rendering of VCMs

Identifying transcription factor targets in neuronal circuits

Prof. McCabe – Group UPMCCABE

©Jamshid Asadzadehatabrizi

Human and animal brains are constructed from webs of interconnected neurons assembled to form neuronal circuits of increasing hierarchical complexity. Changes during development, from experience or in response to disease is thought to alter the function and connectivity of neuronal circuits through modification of the expression of genes within their constituent neurons. However, identifying gene expression changes within circuit neurons has been critically hampered by their intertwined complexity.

In this project, I will deploy new neurogenetic and biochemical technologies designed to isolate and identify gene expression changes within neuronal circuits. I will biochemically purify evolutionarily conserved gene regulatory proteins, known as Transcription Factors, from specific neurons within circuits in a genetic model organism. I will then identify the genes regulated by these factors and determine how their expression is altered in response to changes in circuit activity. My goal is to assemble a genetic portrait of neuronal circuits as they adapt to changes in brain function.

(Start date of fellowship: 1 March 2016)

©Jamshid Asadzadehatabrizi
In this project we aim to identify genomic targets of an established set of transcription factors in neuronal circuits.

New frontiers of discovery and exploration in computational materials science and condensed matter physics

Laboratory of theory and simulation of materials (THEOS)

See publications

Materials science is at the heart of technological innovation. New materials with unusual properties enable new technologies: from super-fast computer memory and smartphone screens to batteries and jet engines. And yet, the number of materials that have been manufactured and characterized experimentally represent a tiny fraction of all materials. This is because creating and measuring new materials in the laboratory is slow and expensive. With the vast treasure trove of revolutionary materials properties lying undiscovered, scientists would like to quickly predict the properties of materials using computers. This research project will create new algorithms to better enable scientists to achieve this aim.

Materials are made up of atoms. To predict the properties of a material we require a very accurate description of how atoms interact (normally based on quantum mechanics), and a very efficient algorithm to explore possible arrangements of atoms in the material. This project is concerned with the second of these two challenges: exploring the arrangements of atoms. The new algorithms we will develop will enable the calculation of materials’ properties at different temperatures and pressures, in a highly automated fashion.

(Start date of fellowship: 1 October 2016)

©Robert Baldock
Predicting the phases of a material at all temperatures and pressures. Heat capacity (Cp) plotted as a function of temperature (T) and pressure (P). Peaks in the heat capacity indicate where the material changes phase from solid (low temperature) to liquid (medium temperatures and pressures) and gas (high temperature).

Novel approaches towards the structure-based design of immunogens for malaria vaccine development

Laboratory of Protein Design & Immunoengineering (LPDI)

See publications

©Jaume Bonet
Vaccines have been an effective instrument to control many infectious diseases. Strikingly, a number of pathogens still lack clinically approved vaccines, generally meaning that classical vaccine development strategies have failed. A new vaccine development strategy has emerged: reverse vaccinology. It relies on the identification of neutralizing antibodies, structural characterization of their target epitopes, and design of carrier proteins able to stabilize and present those epitopes. Currently, the applicability of this approach is limited by the protein scaffolds found in nature onto which the epitopes can be transplanted.

This project aims to develop an innovative computational method to design optimized immunogens by tailoring protein scaffolds around neutralization epitopes (Fold For Function – F3). Creating scaffolds on demand in a systematical fashion represents the next frontier in the field of de novo protein design. F3 will be used as the in silico engine for the design of structure-based epitope-focused immunogens. A central aim in our research is to implement a new Immunogen Design Pipeline, F3 will be in the foundation of this effort which entails a cross-disciplinary effort composed by computational modeling, in vitro biochemical characterization and in vivo immunogenicity analysis.

As a proof of concept, we will perform our methodological developments on the design of immunogens that will mimic key epitopes of Plasmodium falciparum, on the main pathogens responsible for malaria.

(Start date of fellowship: 1 April 2016)

©Prof. Bruno Correia
The Immunogen Design Pipeline (IDP). A comprehensive workflow for the design and characterization of epiope-focused immunogens.


Laboratory of Protein Engineering (LIP)

See publications

©Koerber Stiftung and David Ausserhoefer

Metabolism sustains life. While we now have a good understanding of what key components are essential and how these are chemically converted into each other, we still lack knowledge about their occurrence and transformations in space and time. This project aims to tackle these questions by developing semisynthetic sensor proteins that can sense small metabolic molecules to be visualized in live cells.

The Johnsson laboratory uses self-labeling protein tags that can be endowed with fluorophores and ligands to assess binding events via FRET. The development of such sensors for metabolites and the evaluation of those in primary cells and cell lines is the entry point of this project. Utilizing fluorescence microscopy and fluorogenic probes, the spatiotemporal distribution of metabolites in cells and in their cellular compartments (i.e. ER, mitochondria, golgi, etc.) can be tackled in response to first messengers, drug treatment, ageing and exercise.

(Start date of fellowship: 1 March 2016)

©Johannes Broichhagen
Two self labeling-tags can be orthogonally equippied with fluorophores and a ligand, which leads to high FRET in absence (left) and low FRET in the presence of a metabolite (right).

Computational design of novel 2d-materials for electronic applications

Laboratory of theory and simulation of materials (THEOS)

See publications

©Davide Campi

Novel materials underpin much of the future progress in information-and-communications technologies (ICT) and in energy harvesting, conversion, and storage. 2D materials present an entire novel paradigm and toolbox for materials scientist to discover new properties and functionalities.In absolute terms however, the handful of 2D materials intensively studied up to now represent only the tip of an iceberg.

In this project we want to develop and deploy state-of-the-art first-principles simulations and high-throughput computational workflows to rapidly identify novel 2D materials for breakthrough applications in ICT and energy. Our work start from a massive screening of all known 3D materials to identify exfoliable 2D layers from their parent structures. The most promising candidates (estimated in a few thousands, from an dataset of 200,000 bulk parents) will be subject to refinement on the basis of their mechanical, thermodynamic, and chemical stability, and their optical, electronic, and chemical properties.

(Start date of fellowship: 1 March 2016)

©Davide Campi
Schematic representation of the automatized selection procedure for the identification of the best candidates for nano-electronic applications.


Exploring beige adipocyte origin and differentiation using single cell transcriptomics

Prof. Deplancke – Group UPDEPLA

See publications

©Wanze Chen

Metabolic diseases linked to excessive fat accumulation affect an increasingly large part of the human world population. When one talks about fat, one normally refers to white adipocytes. These cells function to store lipids and constitute as such the largest proportion of human fat. Recent studies revealed that there are other types of fat cells though such as brown adipocytes those, rather than storing fat, burn it to produce heat. These brown adipocytes are abundant in newborns, but diminish over time, so their therapeutic potential seems limited.

However, a third fat type was recently discovered, namely “beige” adipocytes. These beige fat cells appear in certain conditions (e.g., upon cold exposure or high physical activity) in the white adipose tissue, and behave much like brown adipocytes. Their inducible and thermogenic nature makes beige fat cells a great putative target of clinical interventions aimed at regulating energy balance. However, our knowledge of their molecular characteristics, origin, and differentiation is still poor.

The project I propose here aims to identify beige adipocyte precursors, their molecular characteristics and differentiation mechanisms. I will use a multidisciplinary strategy that has at its core the profiling of gene expression in single cells to achieve these goals.

(Start date of fellowship: 1 August 2016)

©Wanze Chen
Strategy to identify beige adipocyte precursors.

A general framework for structure-based de novo design of protein binders

Laboratory of Protein Design & Immunoengineering (LPDI)

©Celeste Hodge

The rational and structure-based design of protein binders to engage relevant cellular targets is of major importance for the study of fundamental biology and biomedicine. However, current methodologies are unable to reliably and robustly design specific protein binders to engage new cellular targets.

Here, we propose a general hybrid computational/experimental platform to de novo design protein binders. This platform will have broad biomedical applications, which we will validate by designing binders to target immune checkpoint proteins for cancer immunotherapy. Inhibiting immune checkpoints is one of the most promising approaches to activate the immune system against tumor cells and our approach could accelerate the development of new therapies.

We will first use an in silico approach to computationally design de novo protein binders. Next, we will characterize the designs biophysically and biochemically and if necessary optimize their properties using directed evolution. Finally, in a collaborative effort, we will evaluate the biological activity of the designed proteins in cellular assays of co-cultured immune and tumor cells.

We foresee that the proposed work could result in new checkpoint inhibitors and that the general methodological framework will be useful for many other important biological and biomedical problems.

(Start date of fellowship: 1 April 2016)

©Pablo Gainza
Computational design and experimental validation of checkpoint inhibitors for cancer immunotherapy.

CryoNanoSIMS to study metabolic transport pathways in dinoflagellates and sea-anemones

Laboratory for Biological Geochemistry (LGB)

See publications

©Louise Helene Søgaard Jensen
Emma Gibbin

Corals thrive in very nutrient poor (sub-) tropical waters because of an endo-symbiotic relationship with photosynthesizing dinoflagellate of the genus Symbiodinium. The metabolic interactions between the coral host and these algae are, however, very difficult to study at the cellular level because most of the nutrients and photosynthates that are exchanged are soluble and are therefore usually lost or severely displaced during sample preparation.

Now, with the recent development of the CryoNanoSIMS at EPFL, direct imaging of the transport pathways of dissolved inorganic carbon (DIC) and nitrogen (DIN), which are fundamental to the metabolism of this symbiosis, has become possible. Using sea-anemones (that harbor the similar dinoflagellate symbionts) a model system (because they are closely related to corals but easier to handle experimentally), I intend to use CryoNanoSIMS imaging, combined with stable isotopic labeling pulse-chase experiments and targeted mechanistic inhibitors to take the very first steps towards revealing the DIC and DIN pathways in the symbiosis without any sample preparation artifacts.

This work will be truly pioneering and will lead the way towards a better mechanistic understanding of how coral reefs will respond to global climate change, including increased sea-surface temperatures and ocean acidification.

(Start date of fellowship: 1 March 2016)

©Louise Helene Søgaard Jensen
The CryoNanoSIMS enables high-resolution images to be taken of cells without the loss, or disruption of- soluble compounds caused by standard sample preparation.

Microstructure embedded white matter tractography of the human brain

Signal Processing Laboratory 5 (LTS5)

See publications

©Gabriel Girard

Diffusion-weighted (DW) magnetic resonance imaging (MRI) is a unique imaging technique because of its sensitivity to the movement of water molecules in biological tissues. By characterizing this movement, it is possible to reconstruct the major neuronal pathways of the brain by means of so-called tractography algorithms. This ability to observe the wiring of the central nervous system allows, for instance, to monitor the recovery after a stroke event or the progress of neurodegenerative diseases.

Although DW-MRI tractography offers an exquisite tool to investigate the architecture of the brain, the relationship between the reconstructed neuronal pathways and the underlying brain’s microstructure characteristics remains poorly understood. Consequently, existing tractography algorithms reconstructions are only indirectly related to these characteristics.

In this project, we want to extend tractography algorithms to reconstruct neuronal pathways while simultaneously estimating the brain’s microstructure characteristics. To achieve this, we will take full advantage of the recent developments in DW-MRI and incorporate microstructure information into the tractography process. This will enable the tractography to distinguishing neuronal pathways with different microstructure characteristics in complex regions of the brain, hence improving the estimation of the wiring of the central nervous system.

(Start date of fellowship: 1 July 2016)

©Gabriel Girard
Brain’s wiring reconstruction using tractography, viewed from above the head. Each line represents an approximation of a neuronal pathway.


Rigid oligoyne surfactants

Laboratory of Macromolecular and Organic Materials (LMOM)

See publications

©Dagmar Rickert

Conventional surfactants have served for long time as model systems for the investigation of the dominant attractive and repulsive intermolecular forces that govern self-assembly of amphiphilic molecules in water. However, decisive aspects of aqueous self-assembly have remained unclear. Particularly, the detailed thermodynamics, that is the complex interplay between enthalpic and entropic changes accompanying self-assembly, are still poorly understood.

This project aims to address the impact of conformational factors on self-assembly of surfactants in water. For this purpose, novel amphiphiles will be synthesized that contain oligoyne segments (conjugated carbon-carbon triple bonds) of different length. In this way, we will obtain a series of surfactants comparable in geometry and intermolecular interactions but exhibiting a systematic variation in rigidity, which accounts for the elucidation of impact of chain entropy on the overall self-assembly.

A comparative investigation of the organization of these amphiphiles in water and at interfaces, along with a detailed thermodynamic analysis of their formation, will be conducted. The proposed work will thus provide new perspectives on the fundamental principles of supramolecular self-assembly in water. Ultimately, we intend to use the as-formed self-assemblies as soft template for the creation of novel carbon nanomembranes, since the oligoyne segments can be cross-linked by UV-irradiation.

(Start date of fellowship: 1 March 2016)

©LMOM group
Schematic illustration of the self-assembly of surfactants with differently long (flexible) alkyl and (rigid) oligoyne segments and, consequently, a systematic variation of their conformational freedom in water and at interfaces.


Fundamental physics behind high-efficient hybrid perovksite solar cells


See publications

©Giulia Grancini

In front of the increasing global energy demand, solar energy will be a major player in the energy market. However, a cost effective -in terms of materials and processing- solar technology is needed. Stemming from solution-processable semiconductors, a new concept based on hybrid semiconductor (methyl ammonium-lead iodide perovskite) is challenging the photovoltaic community. This nascent technology has already grown at impressive rates demonstrating power conversion efficiency beyond 22%. Behind huge research efforts, the optimization protocols have been so far driven too much by serendipitous improvements that sometimes lack of a rational understanding and scientific reproducibility.

To enable a real progress beyond the state of the art the project aims to set a new strategy for an educated device design driven by a fundamental comprehensive understanding of the main physical processes behind solar cell operation. This throughout analysis will define the inextricable relationship between structural, optical and photophysical properties behind established and new perovskites and reveal the fundamental processes behind device operation (i.e. charge injection, transport and recombination) aiming at high efficient and stable solar cells.

This interdisciplinary approach will lead to successfully achieve the results producing a significant advances in the fundamental knowledge generated and high technological impact with benefits for Swiss economy.

(Start date of fellowship: 1 March 2016)

©Giulia Grancini
a.Cartoon of a photovoltaic cell based on perovskite active material sandwiched between electron and hole transporting layer. The photoinduced processes such as electron and hole transfer at the interfaces will be investigated to improve the device performances. As an example, in b., the photoluminescence decay upon light absorption in the perovskite will be monitored.

Sparse-view differential phase contrast computed tomography using low-rank fourier interpolator

Biomedical Imaging Laboratory (LIB)

See publications

©Yo Seob Han

The aim of this proposal is to develop a new algorithm for differential phase contrast computed tomography (CT), which can maintain the quality of reconstructed tomographic images while reducing scanning views leading to short time exposure to x-ray. Currently, the state-of-the-art techniques addressing sparse-view or low-dose CT are formulated as penalized least square or maximum a posteriori (MAP).

Despite these recent advances, features such as complex texture, small structures, or outlines of organs, which are significant to diagnose diseases, are not satisfactorily reconstructed with current techniques in the extreme scenario of limited angles or significant noise contamination. Such systematic errors originate from the limited coverage of regularization seeking the total variational sparsity, edge-preserving, or wavelet transform, etc.

In order to address such limitations, we shall explore the fundamental relationship between sparse stochastic processes (SSP) model and annihilating based low-rank Hankel matrix approach (ALOHA). The SSP model enables us to solve linear inverse problems by formulating MAP estimators derived from sparse stochastic processes. The ALOHA approach, on the other hand, exploits the annihilation property of transformed signals via Hankel matrix completion for ill-posed inverse problems. Our proposal here is to develop a better reconstruction method by combining those two approaches.

(Start date of fellowship: 1 June 2016)

©Kyong Hwan Jin
Major goals of this project, and three components of the proposed algorithm for sparse-view differential phase contrast x-ray computed tomography (CT) reconstruction.

Turning nitrous oxide a real oxidant with Cu complexes

Laboratory of Supramolecular Chemistry (LCS)

See publications

©Kay Severin

While carbon dioxide (CO2) has drawn considerable attention due to its environmental impact in recent years, a molecularly relevant species, nitrous oxide (N2O), deserves no less public awareness. N2O has now been identified as the most potent ozone-depleting substance emitted in various industrial and agricultural processes. It is also a very effective greenhouse gas, with approximately 300 times greater warming potential than CO2.

On the other hand, N2O is thermodynamically a very strong oxidant, and its only byproduct, nitrogen gas (N2), is environmentally benign. This underlines the promise of utilizing N2O for producing value-added chemicals, which is however under-developed due to the low reactivity of N2O.

In Nature, copper-containing proteins are employed to catalyze the transformation of N2O to N2. Following this line, I will design a series of copper complexes and investigate their ability in activating N2O. This will help us understand the key structural characteristics of these catalyst models for fulfilling such a purpose. As a proof of concept, I will then apply the promising molecules to producing industrially useful chemicals such as organic amines and epoxides where N2O is utilized as the sole oxidant.

(Start date of fellowship: 1 March 2016)

©Yizhu Liu
Figure 1 Development of Cu complexes as catalysts for oxidizing organic substrates using N2O as the oxidant.

Microengineering the anterior-posterior patterning of the human epiblast

Prof. Lutolf – Group UPLUT

See publications

©Alice Lucchin

How is the body plan of a living being established? During embryonic development, cells of the epiblast differentiate in a controlled manner so that, starting from a uniform population of cells, the extremely complex but organized structure of the body arises in a process called patterning.

Studies on animal models suggest that this phenomenon is largely controlled by concentration gradients of key signaling molecules. According to its position in the epiblast, a cell will be subjected to a specific concentration of signaling molecules, which in turn will translate in the acquisition of a specific cell fate. The possibility to directly study in vivo how signaling molecules spatially influence cell fate decisions in the epiblast is far from being trivial, but next-generation in vitro systems, in combination with human Pluripotent Stem Cells (hPSC), could fill this gap.

The goal of this project is to build an in vitro microsystem that, by exposing hPSC to spatially oriented gradients of signaling molecules, could recapitulate specific aspects of epiblast patterning. This will allow to obtain new insights on how gradients of signaling molecules influence developmental processes and more specifically the spatial patterning of cell fate.

(Start date of fellowship: 1 May 2016)

©Andrea Manfrin
Top row: Schematic of the experimental procedure. Bottom row from left to right: representative picture of the fluidic setup; experimental validation of the generation of gradients in the microfluidic device; hPSC pattern of mesendoderm after exposure to gradients of signaling molecules.

The role of Mitofusin-2 in vulnerability to stress induced depression

Behavioral Genetics Laboratory (LGC)

See publications

©Borja Rodríguez-Herreros

Chronic exposure to stressful life events increases the risk to develop psychopathologies, including depression. However, the mechanisms whereby some individuals become depressed after stressful experiences whereas others can cope with stress easily remains unknown. The mitochondrial hypothesis of stress vulnerability to depression proposes that individuals with impaired mitochondrial function have lower capacity to cope with stress, rendering them more susceptible to develop depression.

In this project, we aim to investigate whether the expression of the mitochondrial protein Mitofusin 2 (MFN2) in the mesolimbic system –a key brain region linking stress with some depression symptoms- regulates the susceptibility of individuals to develop stress-induced depression. MFN2 exerts a multifaceted control on mitochondrial function, playing roles in mitochondrial fusion, mitochondrial respiration and on the transport of mitochondria from soma to synapses, all of them critical processes to meet the energy demands required by essential neuronal processes, such as neurotransmission. The results from this project may advance the understanding of the mechanisms involved in individual differences in vulnerability to develop depression and pave the way for the discovery of novel treatments.

(Start date of fellowship: 1 April 2016)

©Laia Morató and
This project hypothesizes that chronic stress decreases MFN2 expression in the mesolimbic system of susceptible mice. This decrease would cause alterations in mitochondrial function (e.g. fusion, respiration, transport), which would play a critical role in the development of stress-induced depression.

Catalytic atropoenantioselective synthesis of biaryls via CH functionalization

Laboratory of Asymmetric Catalysis and Synthesis (LCSA)

See publications

©Michael Sherburn

Most pharmaceuticals, agrochemicals, as well as many valuable material and commercial products are synthesized in organic chemistry laboratories, thus the development of new reactions, strategies, techniques, or improvements in reaction selectivity, efficiency and economy are of great importance. In many of these applications the ability to access to a single enantiomer of a molecule is a necessity. Enantiomers are a pair of molecules that are non-superimposable mirror images, and as a result of their structural similarity the selective synthesis of one can be a challenging task.

We are interested in the development of new catalytic enantioselective CH functionalization reactions. This class of reaction has several advantages over more traditional cross-coupling alternatives in that substrates do not require organometallic pre-functionalization, are often less toxic, air and moisture sensitive, and their application generates less waste. We are looking at applying these new techniques to the synthesis of biaryl atropisomers, a special class of enantiomer that has to be prepared with care as they can convert into their mirror image at elevated temperatures. This structural motif is present in many biologically important molecules, and acts as the stereochemical controlling element in many ligand scaffolds.

(Start date of fellowship: 1 March 2016)

©Christopher Newton
The development of new enantioselective CH functionalization reactions and their application to the synthesis of atropisomers.

Deciphering the roles of transposable elements in 3D genome architecture and gene expression

Laboratory of Virology and Genetics (LVG)

See publications

©Julien Pontis

Establishing and maintaining the pluripotency of embryonic stem cells and controlling their lineage commitment is critical for the development and homeostasis of higher organisms. Key to this process is the influence of DNA regulatory sequences such as promoters, enhancers and insulators, many of which reside in transposable elements (TEs).

TEs display exquisitely specific patterns of expression both during development and in adult tissues, and their deregulation has been observed in human diseases, for instance some cancers and neurological disorders. However the knowledge of their involvement in gene regulation and the extend of their impact in development and homeostasis is still rudimental. One model system is the human Embryonic Stem Cell (hESC) which represent a highly interesting system for studying the influence of TEs on transcriptional networks, because in these cells TEs are particularly active, partake to the maintenance of pluripotency and are subjected to highly dynamic control.

Accordingly, I will use hESC as a paradigm to study the impact of TEs on the regulation of gene expression. I will tackle two major questions during my postdoc: (i) how TEs influence transcription in hESCs, and (ii) how TEs more broadly contribute to the establishment of human-specific gene regulatory networks. Those immediate aims will raise a solid basis for the understanding of their potential roles in pathological conditions.

(Start date of fellowship: 1 August 2016)

©Julien Pontis
How distal TEs participate to human regulatory transcriptional network.

Innovative public procurement as an innovation policy

Chair of Innovation and IP Policy (IIPP)

See publications

©Horvat Endre (EPFL)

This project aims to provide sound quantitative evidence on the importance of innovative public procurement (IPP) in stimulating innovations that are valuable for firms and for the society as a whole.

IPP is defined as the purchasing activities carried out by public agencies that lead to innovation. During the last decade demand-side innovation policies have received a renewed interest, and IPP has been increasingly considered as a de facto technology policy tool from researchers and governments. Despite the rising interest, the extant literature is silent on: i) whether procurement spurs innovations that are economically valuable; ii) the role played by the size of IPP contracts on the quality of the output of procurement processes; iii) whether IPP is an effective policy to foster structural technological change, especially in green technologies. This project fills these gaps to advance knowledge on IPP.

The analysis combines data from off-line and on-line administrative sources and uses state-of-the-art econometric techniques. The research design entails the construction of an original dataset linking, for the first time, procurement data with patent and firm data. Results have direct policy implications and help policy makers to tailor effective technology policies.

(Start date of fellowship: 1 August 2016)

Public procurement may spur technological change and have an important role in addressing grand societal challenges, such as the transition to climate-friendly technologies.

Plasticity for normal contact between rough surfaces

Computational Solid Mechanics Laboratory (LSMS)

See publications

©Valentine Rey

Contact and friction are pervasive in many natural phenomena and engineering applications. When studying earthquakes or the evolution of fretting damage in materials, a detailed comprehension of the physical mechanisms at contact interfaces is of fundamental importance.

A major challenge in contact mechanics stems from the multi-scale nature of phenomena occurring at the contact: natural surfaces are rough which means they are made of a vast number of asperities over a wide range of length scales. Therefore, when two rough surfaces are in contact, the real contact area is much smaller than the apparent contact area.

Accurate quantitative results for frictionless elastic contact between large rough surfaces can be obtained through numerical simulations. The calculated pressures at micro contacts are so high that non-linearity is expected to occur at these micro contacts. In the case of metallic materials, the asperities are likely to deform irreversibly, which modifies the overall contact response.

The fundamental objective of the project is to account for non-linear mechanical behavior between rough surfaces within an efficient contact algorithm. Statistical expansions will thus be proposed to give simple expressions of the evolution of the contact area with pressure, with direct applications in engineering.

(Start date of fellowship: 1 April 2016)

©Valentine Rey
Modelisation of a fractal rough surface.

Immune-tolerance and the hematopoietic stem cell niche: redefining the mensenchymal to preadipocyte differentiation axis within the bone marrow

Naveiras Group – GR-NAVEIRAS

©Shanti Rojas-Sutterlin

Hematopoietic stem cells (HSCs) produce all blood cells. They reside in specialized areas of the bone marrow (BM) called niches where they interact with perivascular mesenchymal stem cells (MSCs), which ensure HSC survival and proliferation. When HSC function is impaired, severe BM failure installs and this may be caused by toxic or autoimmune insults.

At the cellular level, BM failure is characterised by massive infiltration of mature adipocytes, a process known as red-to-yellow transition. It was shown that mature adipocytes inhibit, whereas adipocyte progenitors, including MSCs and preadipocytes, support hematopoiesis in the mouse. Interestingly, the yellow-to-red transition can be manipulated and mice treated with adipocytic inhibitors show accelerated hematopoietic recovery. These results support the importance of the adipocytic differentiation axis for hematopoietic function. However, the MSC/preadipocyte subpopulations with hematopoietic-supportive or immunomodulatory properties accumulating upon adipocytic inhibition in BM failure have not yet been characterized. Furthermore, the validity of the above-mentioned findings has not been addressed in human.

The purpose of this study is to identify and characterize MSC/preadipocyte populations with hematopoietic-supportive and immunomodulation potentials in human tissues. I believe that targeting the HSC niche to accelerate hematopoiesis recovery could be complementary to the current treatment that directly targets HSCs.

(Start date of fellowship: 1 March 2016)

©Shanti Rojas-Sutterlin and Olaia Naveiras
Schematic representation of the approach to accelerate hematopoietic recovery upon BM failure. (Tx: transplantation)

Perovskite/silicon tandem for sustainable energy production

Laboratory of Photonics and Interfaces (LPI)

See publications

©Manuel Adams

Global energy production more than doubled in the last four decades, with future projections indicating an ever increasing demand for more energy. Currently, over 80% of the global energy production are from fossil fuels that produce massive carbon emissions causing climate change.). On the other hand, the main source of energy in the solar system, the sun, is free of cost and abundant. Only one hour of sun light (received globally) is sufficient to power the entire planet for one year. Solar cells tap into this vast potential and are therefore key towards a long-term energy solution with a low carbon footprint.

Currently, silicon solar cells are by far the most important and relevant solar energy source with a globally established infrastructure and distribution network. However, further cost reductions are more challenging to achieve as the practical maximum is reached and small improvements come at disproportionately high costs.

In the past 5 years, perovskite solar cells have emerged with impressive efficiency gains. Interestingly, currently used perovskites have a band gap which can complement established silicon solar cells. Therefore, instead of competing, silicon and perovskite solar cells can be combined in a so-called “tandem” architecture uniting the unique advantages of both.

(Start date of fellowship: 1 March 2016)

©EES journal (!divAbstract)
Schematics of a highly efficient monolithic perovskite (top) / silicon (bottom) tandem solar cell that unites the unique advantages of both technologies.

Probing exosome biogenesis and uptake by using anthrax toxin as a tool

Prof. van der Goot Group – VDG

See publications

©Sylvia Ho

Many cells use vesicles to communicate with each other. One type of these vesicles, exosomes, are derived from endosomal compartments in the primary cells and then taken up by naïve cells. They have been implicated in a variety of diseases, mainly cancers. However, understanding how exosomes are made from cells, let alone from diseased cells, has been a challenge.

Recently, it was discovered that lethal factor (LF) of anthrax toxin can be transferred to naïve cells via exosomes. LF makes its way into the cytoplasm, where it can cleave MAPKKs. Using anthrax toxin as a tool allows us to observe and probe exosome biogenesis and uptake without purification of the exosomes, a big issue in exosome research. We have developed an siRNA screen to find genes that may be implicated in the exosome pathway and are in the process of screening more than a thousand genes.

Candidate-based approaches have already revealed a few interesting candidate, including that exosome uptake is mediated by GPI-anchored proteins, and that ER stress inhibits both exosome release and uptake. We are further probing these mechanisms and trying to understand the molecular interactions involved. We hope to gain a general understanding of exosome biology via our unique system.

(Start date of fellowship: 1 March 2016)

©Oksana Sergeeva
Figure showing entry of anthrax toxin (PA and LF) into the primary cell, and then secretion of intraluminal vesicles (ILVs) as exosomes, and their uptake by naïve cells.

Identification of novel compounds to treat rare mitochondrial diseases

Nestlé Chair in Energy Metabolism (NCEM)

See publications

©Vincenzo Sorrentino

Mitochondria are the energy centers of the cell, using fuels (glucose, fat, etc) to produce energy (heat or ATP, the energy currency of the cell). Impairment of these organelles, caused by mutations in the genes essential for the production of energy, is the cause of about 150 rare metabolic and neurological diseases. Mitochondrial function can be improved by activating or boosting two main quality control processes in the cell: 1) mitochondrial biogenesis, which leads to the formation of new, more functional mitochondria, and 2) UPRmt, a mitochondrial repair response activated by stressors which can damage mitochondria.

Pharmacologically targeting mitochondrial biogenesis and/or the UPRmt pathway hence may improve mitochondrial dysfunction and favorably impact on mitochondrial diseases. Therefore, the goal of this research project is to identify novel compounds that increase mitochondrial biogenesis and activate UPRmt using a multispecies and multi-scalar screening strategy in mammalian cells and in C. elegans. Compounds with drug-like features that induce biogenesis/UPRmt will be subsequently tested for their efficacy to improve mitochondrial dysfunction in pre-clinical cell, worm and mouse models of mitochondrial disease. Our screening and subsequent validation of our top candidates will pave the way towards the development of new and efficient drugs for mitochondrial diseases.

(Start date of fellowship: 1 April 2016)

©Vincenzo Sorrentino
Dysfunctional mitochondria can lead to metabolic impairment and disease. A drug screening will be performed with the aim to identify compounds that boost mitochondrial biogenesis and the UPRmt repair pathways to restore mitochondrial function and treat mitochondria diseases.

Dynamic cortical integration in perceptiondependent behavior revealed by wide-field voltage and calcium imaging combined with optogenetic intervention in task-performing mice

Sensory Processing Laboratory (LSENS)

©Mayuko Tamura

Imagine that you pick up an object. You perceive what (shape, pattern) and where (place, movement) the object is, and then prepare and execute motor commands for contracting and relaxing muscles of arms, hands and fingers. In your brain during this simple behavior, you process a large amount of information parallelly and serially by coordinating diverse neural activities in different areas of the brain. Global integration of local processing is thus thought to be a fundamental principle of the brain, which even enabled high-level cognitive behaviour in humans.

In local brain areas, unique properties of various types of neural cells have been revealed in recent studies. However, mechanisms in which such local activities are integrated into the brain-wide processing remains unclear. In this project, we focus on spatio-temporally-selective actions of different types of neural cells in the whole neocortex of the brain. In mice performing a whisker-stimulation detection task, we combine wide-field optical imaging and optical manipulation of cell-type specific activities. We will uncover what types of cells are recruited, where in the cortex, when in the task periods, and how these cells act on communication among cortical areas to process whisker perception and produce required behavioural responses.

(Start date of fellowship: 1 November 2016)

©Keita Tamura
Spatio-temporally selective actions of different types of neural cells underlying global integration of diverse local processing, which produce appropriate behavioural outputs.

Active plasmonic metasurfaces for mid-infrared phase control and adaptive optics

Bionanophotonic Systems Laboratory (BIOS)

See publications

©Andreas Tittl

Optical imaging devices in the mid-infrared play an important role for many technological applications in areas such as thermography, automotive safety, and astronomy. This functionality relies on the atmospheric transmission windows in the 3–5 μm and 8–12 μm spectral ranges, which allow such devices to work without disturbance from water vapor, dust, or other atmospheric influences.

One key challenge in this range is the experimental realization of an optical device with spatial control over both the amplitude and the phase of incoming light (a spatial light modulators or SLM), which is not readily available in the mid-IR.

In this project, we will utilize a plasmonic metasurface consisting of resonant subwavelength nanoantennas to manipulate light on the nanoscale. We can then combine this metasurface with an active material layer to gain external control over the optical properties. As a final step, the active metasurface can be subdivided into individual pixels to obtain the desired spatial control over the light transmission.

We believe that our approach will allow for the technological realization of ultra-thin amplitude and phase modulators with a wide array of applications ranging from adaptive optics for infrared astronomy and night vision systems to improved free-space optical communication systems.

(Start date of fellowship: 1 August 2016)

©Andreas Tittl
An active plasmonic metasurface arranged in a pixelated pattern can realize efficient modulation of mid-infrared radiation in an ultra-compact design.


Spatiotemoral investigation of ultrafast dynamics of low-dimensional nanomaterials driven by coherent control of lattice and electronic degrees of freedom

Laboratory for ultrafast microscopy and electron scattering (LUMES)

See publications

©Giovanni Maria Vanacore

The development of innovative approaches for energy conversion and storage represents one of the main scientific challenges of the 21st century. The emergence of two-dimensional (2D) materials has the potential to decisively address these issues. In these systems, made of a single-layer of atoms, the quantum confinement of electrons, photons and phonons in a 2D space gives rise to exciting new properties that are not present in their 3D bulk counterpart. An enormous effort is being devoted toward their comprehension and improvement, however, to fully exploit their potential it is crucial to be able to control and tailor their physical properties.

To achieve this objective, we propose to use electromagnetic waves with THz frequency to coherently manipulate the atomic and electronic structure of 2D materials. The THz excitation will allow to resonantly excite specific degrees of freedom, inducing subtle modulations that can lead to significant changes of the material properties. To visualize this transient nonequilibrium behavior, we will adopt ultrafast electron diffraction and imaging, which can provide simultaneous observation of their dynamics in space and time with the required atomic sensitivity and femtosecond resolution.

The proposed approach is not only very intriguing from a fundamental perspective, but is also a crucial pre-requisite for achieving a real breakthrough in energy-based applications in the near future.

(Start date of fellowship: 1 March 2016)

©Giovanni Maria Vanacore
Schematic representation of the proposed experimental approach for coherent manipulation of 2D materials and their visualization in space and time. An ultrafast THz pulse is used to selectively excite atomic, spin or electronic degrees of freedom, and an ultrashort electron pulse is able to map the nonequilibrium atomic scale dynamics of the system.


From spikes to electrocorticograms: brain-spinal interfaces across cortical spatial scales to restore locomotion after spinal cord Injury

IRP Chair on Spinal Cord Repair (UPCOURTINE)

See publications

©Fabien Wagner

Spinal cord injury leads to permanent locomotor impairments that seriously diminish the patients’ quality of life. An electronic interface between brain modulation and electrical spinal cord stimulation has the potential to bypass the injury, and thus to restore brain control over leg movements. However, the most relevant type of gait-related brain signal to implement this brain-spinal interface remains unknown.

Here, I will compare the suitability of neuronal signals recorded at the surface of the cortex (micro-electrocorticograms, or μ-ECoG) versus in the deep layers of the cortex (intracortical activity) to control selective spinal cord stimulation protocols in freely behaving non-human primates. These signals present specific trade-offs between spatial resolution, surgical invasiveness and long-term stability. These factors play a critical role for the safe and efficacious implementation of a brain spinal interface in clinical settings.

Macaque monkeys will be implanted with intracortical microelectrode arrays and soft surface microelectrode implants to record intracortical versus electrocorticogram activity, and a spinal multi-electrode implant over lumbar segments of the spinal cord to deliver electrical stimulation. I will integrate these technologies to evaluate the long-term performance of brain spinal interfaces based on intracortical versus electrocorticogram signals to adjust leg movements during locomotion in healthy monkeys.

(Start date of fellowship: 1 May 2016)

©Courtine lab.
Figure 1: Brain-Spinal Interfaces across cortical spatial scales. A combination of wireless technologies with real-time capabilities enables to link decoded motor intentions from motor cortex signals to adjustment of electrical spinal cord stimulation in monkeys freely moving on a treadmill or overground. Recordings can be based either on intracortical arrays or μ-ECoGs fabricated with the e-dura technology (developed by Prof. Stéphanie Lacour at EPFL).


Rethinking myoelectric control: an exploration of recording methods and novel control strategies

Bertarelli Foundation Chair in Translational Neuroengineering (TNE)

See publications

©Katie Zhuang

One of the most active fields of research in rehabilitation is in the design of myoelectric prostheses, which aim to restore natural movement functions to those who have lost it. These allow users, usually amputees, to control prosthetic devices with residual muscle activity. However, user control strategies in commercial products lag far behind the development of the prosthetic devices themselves.

First, the various strategies to obtain myoelectric recordings can be either unreliable or impractical for widespread clinical use. Next, algorithms for decoding peripheral nerve signals have improved much over recent years, but still suffer from misclassifications and prediction of unwanted movements. Finally, most myoelectric prostheses impose a high cognitive load and require the user’s visual attention for robust control.

To address these problems, we propose to compare costs and benefits of existing myoelectric recording strategies, develop a novel training paradigm that will cater to the capabilities of each individual user and lastly, introduce a shared-control scheme in which precise movement kinematics are automated. The analysis of the improvement of performance using (invasive and non-invasive) sensory feedback will be also investigated. Results of this project can directly advance the field of neuroprosthetics and guide the next generation of myoelectric devices.

(Start date of fellowship: 1 May 2016)

©Katie Zhuang
A neuroprosthetic device controlled by a user’s muscle activity.