Selected Fellows Call 2014

Below you will find an overview of the proposals of the successful candidates of the 2014 call by alphabetical order.

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

The unconventional roles of calnexin in signalling and apoptosis

Van Der Goot Lab (VDG)

See publications

©Nicole Andenmatten


The endoplasmic reticulum (ER) is a large, membrane-bound organelle in eukaryotic cells that fulfils key functions including the folding and quality control of more than 1/3 of all cellular proteins.

Interestingly, the Van Der Goot lab has recently found that cells integrate information on the ER health status to tune the cellular response to certain external stimulus. More specifically, the transmembrane ER chaperone calnexin appears to act as an important decision making point in the cell that controls cell proliferation and apoptosis, two processes central to embryonic development and diseases such as cancer.

While sensing of ER health occurs in the ER lumen, modulation of signalling involves caspase-8 dependent cleavage of a cytosolic domain of calnexin. This calnexin tail-domain subsequently translocate to the nucleus where it interacts with downstream factors thereby affecting gene transcription and cellular faith.

I propose to gain a better understanding on how the calnexin tail can undergo caspase-8 mediated cleavage in response to specific stimuli, how the release of the cytosolic domain is regulated in time and space and most importantly to dissect the yet uncharacterized pathway by which the cytosolic tail inhibits apoptosis in cells devoid of ER stress.

(Fellowship dates: 1 – 31 August 2015)

Project illustration
©Nicole Andenmatten, adapted from Lakkaraju et al 2014
Calnexin and signaling: Calnexin plays a role in sensing the endoplasmic reticulum health status and integrating this information into cellular outcomes of extracellular stimuli (e.g. EGF receptor).

Computer aided identification and study of the nature of performances degrading defects in semiconductor materials

Chair of Atomic Scale Simulation (CSEA)

See publications

©Assil Bouzid

Transistors are the elementary building blocks of modern electronics. Conventional transistors are based on silicon and its derived oxides. This widespread use of silicon is promoted by its low manufacturing cost and easy engineering industry. However, the SiO2 oxide has reached the end of the technological road map and further scaling with this material is prevented since fundamental limits due to quantum tunneling cannot be overcome. Many alternative semiconductors were proposed since in their ideal structures they appear to be superior to silicon. Among them, the III-V group As based compounds attract considerable interest.

However, unlike for silicon, the precise control of the surface/interface properties of III-V materials is hindered by the formation of performance-degrading defects. While the identification of these defects could help in finding the possible passivation routes and improve the device performances, their origin, formation mechanism, and electronic properties are still poorly understood.

In this project I will combine the constant potential grand canonical simulation technique with first principle molecular dynamics simulations, with the aim of establishing a complete computer-aided framework of identification and study of defects in semiconductor materials. This project is intended to fill the longstanding need of a theoretical scheme to assist and complement the worldwide growing effort of engineering surfaces and interfaces.

(Start date of fellowship: 1 January 2016)

Figure: Schematic illustration of the three-step method toward defect identification and study

© Assil Bouzid

Computational materials design of novel perovskites for photocatalytic water splitting and for the oxygen evolution reaction

Theory and Simulation of Materials (THEOS)

©Ivano E. Castelli

The harvesting and storage of solar energy via water conversion into molecular hydrogen, produced in the water splitting reaction, is one of the most promising avenues for an economy based on renewable energies and reduced emissions. The requirements for an efficient photocatalyst are complex: first, one needs a stable material able to absorb a significant part of the solar spectrum; then, once an electron-hole pair has been created, the electron and the hole need to reach the surface and be energetic enough to run the water splitting reaction. An ideal material has also to display optimal thermodynamics and catalytic properties.

In this project, using high-throughput screening techniques, I will expand the search for novel materials for light harvesting into the organometal halide and lanthanides perovskites and I will study the oxygen evolution reaction, that is the bottleneck for using water splitting devices in common life. For example, organometal perovskites are extremely interesting because of their high absorption properties and high charges mobility, but they contain lead which is toxic. A challenge for the scientific community is to find organometal materials lead-free. The final goal is to identify the most promising candidates on which to focus the experimental efforts.

(Fellowship dates: 1 March – 31 August 2015)

High-throughput scheme: I will study perovskite materials to find novel candidates for the oxygen evolution reaction and as light harvester in a photoelectrocatalytic device using ab-initio quantum mechanics simulations.

©Ivano E. Castelli

Generation of high-affinity ligands based on tricyclic peptides for therapeutic application

Laboratory of Therapeutic Proteins and Peptides (LPPT)

See publications

©Kaycie Deyle

Current therapeutics can be classified as small molecules or large biomolecules, but there is a gap between these technologies that should be spanned in order to more successfully treat a variety of diseases.

Peptides, small subsets of larger biomolecules, are now being considered as this potential bridge because they combine desirable target binding traits of biomolecules with the size and ease of synthesis of small molecule drugs. These peptides, however, still suffer from a few major drawbacks, such as relatively low target binding affinity and short half-life in the body, that prevent them from being widely used for drug development.

This project aims to create a new format of chemically-modified peptides designed to attach to a larger biomolecule scaffold which will act as a rigid framework to stabilize the peptides. The scaffold confers a structural constraint that should provide a higher affinity for the drug targets of interest than the floppy, unconstrained peptides. This design will also allow for peptides to be administered as small molecule drugs which will then tether to their scaffolds, which will be biomolecules constantly present in the bloodstream, in order to benefit from the increased half-life of these large molecules.

(Start date of fellowship: 1 June 2015)

Perovskites: new materials for solar energy conversion.
Unconstrained peptides are chemically-modified to attach to a large biomarker found in the blood. This serves to constrain the peptide to increase the affinity for the drug target and the large biomolecule increases the size (half-life) of the therapeutic.

©Kaycie Deyle

Optimizing the magnetic properties of metal-organic nanostructures through surface interaction

Nanostructures at Surfaces (LNS)

See publications

©Katharina Diller

Today’s ongoing computerization leads to a growing demand for non-volatile memory, especially in fields such as telecommunication and cloud computing, as well as scientific simulations. New approaches aim at using (metal-)organic materials to replace the classical semiconductor memory storage devices, in the same way that organic solar cells replace their classical counterparts.

Hereby metal atoms (which provide the necessary spin) are embedded in layers of organic molecules, which act as spacer between the metal atoms. For possible applications it is necessary to stabilize the molecules on the surface of an electrode. In the simplest case this does not alter the functionality of the molecules. More often, however, the molecule-substrate interactions influence the magnetic properties. This might be seen as an additional problem, but can also be used as an advantage, namely as an additional tool to tune functionality by choice of a suitable molecule-substrate system.

For this project different molecule-metal samples will be prepared by depositing the nanometer-sized organo-metallic building blocks under ultra-high vacuum conditions on clean metal surfaces. The systematic comparison aims at improving the general understanding on how the magnetic properties can be tuned by altering the adsorption properties and thus provide a tool-box for tailoring magnetic properties on a molecular level.

(Start date of fellowship: 1 April 2015)

The interaction between metal-organic nanostructures and the substrate can strongly influence the magnetic properties of the system.

©Katharina Diller

Neural and cognitive bases of self-grounded perception

Laboratory of Cognitive Neuroscience (LNCO)

See publications

Nathan Faivre ©Ema Cima
Nathan Faivre

©Ema Cima

For the last twenty years, there have been considerable efforts to describe the neural correlates of perceptual consciousness, that is the minimal set of neuronal events sufficient for a subjective experience to occur.

Another fundamental property of consciousness is its link with the self, that is the subject of consciousness experience. The close link between the self and body representation has been termed bodily self-consciousness and has been a target of recent research, benefiting from the examination of neurological patients and virtual reality coupled with robotics.

In this project, we propose to characterize the fundamental relations between visual and bodily self-consciousness in healthy and clinical populations, at the limb and full-body level, for simple and complex cognitive processes, through behavioral and neural measures.

At the theoretical level, the hypothesis of self-grounded perception is supported by the most influential theories of perceptual consciousness which state that conscious perception involves information integration throughout the brain. It is also backed up by a theoretical account of bodily self-consciousness which considers multisensory integration as a key mechanism for the sense of self. In our view, these multiple conceptual and methodological tools will provide valuable knowledge about the fundamental interplay between perception and the self.

(Start date of fellowship: 1 March 2015)

We propose that the neural encoding of a visual object depends on the sense of self, involving both the TPJ (temporo-parietal junction) and the insula.

©Nathan Faivre

Solition physics in on chip SiN micro-resonators for frequency comb applications

Laboratory of Photonics and Quantum Measurements (K-LAB)

See publications

©Michelle Cherance

By coupling light of a single frequency laser into a micrometer sized resonator (microresonator) fabricated out of a highly nonlinear material, several new laser lines can be generated. When these laser lines are equidistant in frequency they constitute a frequency comb. Microresonator based frequency combs are of special interest to metrology, spectroscopy, telecommunication and laser applications due to their wider frequency spacing between two comb lines, compared to commercially available fiber based frequency combs.

At EPFL we develop and fabricate planar chip based silicon nitride microresonators on silicon wafers, which makes the process scalable in volume and price. For this project we are driving the developed platform towards applications.

In order to achieve this we must first deepen our understanding of the formation of pulses (solitons) in the resonator that constitute the frequency comb. These frequency combs can then be used to create low noise microwave oscillators, widely used in timing and navigation. Furthermore, the microresonator frequency combs offer a high repetition (video) rate at high resolution useful for spectroscopy applications and calibrations. We endeavor to show in-situ near infrared spectroscopy and build a device to calibrate astrophysical spectroscopes in search of exoplanets with highest precision.

(Start date of fellowship: 1 July 2015)

©Victor Brasch – EPFL
Perovskites: new materials for solar energy conversion

Photonic chip with several silicon nitride microresonators that can convert a single frequency to a frequency comb.

Translational study with end‐users on non‐invasive neuroprostheses using cognitive signals and shared‐control schemes

Defitech Chair in Brain-Machine Interface (CNBI)

See publications

©Iñaki Iturrate


Electroencephalography (EEG)-based Brain-machine interfacing (BMI) is an emergent technology developed to provide a communication channel between a human and a device using only brain activity.

Since they do not rely on muscular activity, they can provide a channel of communication and control for users with severe neuromuscular disabilities. Despite EEG-based BMIs have shown successful applications such as communication devices or the control of neuroprosthetic devices such as wheelchairs, non-invasive BMIs lack from out-of-the-lab solutions.

In this project, we will demonstrate how non-invasive BMIs can provide real market-oriented solutions, and how careful control designs allow to close the performance gap between invasive and non-invasive systems. To this end, we will follow two different paths. In the translational path, the objective will be to exploit current state-of-the-art machine learning techniques to design a new BMI-based communication device with a user-centered design that can become a final prototype for the market. In the basic research path, we will focus on the study of EEG correlates of high-level decisions during a reaching and grasping motor task, so far only known to be detectable with invasive recording systems.

(Start date of fellowship: 1 April 2015)

©Iñaki Iturrate
General scheme of the EEG-based communication device proposed.

The role of energy metabolism at the cortico-nucleus accumbens circuit in the biological resiliency and susceptibility to develop depression under stress

Laboratory of Behavioral Genetics (LGC)

See publications

©Rami Rishmawi

Sustained stress is closely linked to the onset and symptom severity of common affective disorders such as depression. Interestingly, a proportion of humans exposed to adversity does not show any signs of psychopathology and maintains physiological stability. Despite its importance, little is known about the neurobiological mechanisms that mediate vulnerability or resistance to chronic stress.

Several lines of evidence suggest a role for mitochondrial abnormalities in stress-related depression. Indeed, the brain has a high energy demand and must be able to maintain normal function in response to environmental challenges through mitochondrial energy production.

Hence, it has been suggested that enhancing mitochondrial function is related to increased adaptability and biological resilience. This led us to an intriguing question regarding whether innate mitochondrial function differences and/or mitochondrial susceptibility to chronic stress-induced changes can causally contribute to the development of depression-like symptoms induced by exposure to chronic stress in mice.

(Start date of fellowship: 1 March 2015)

Fig. In this proposal we pursue an intriguing research question namely whether innate mitochondrial function differences and chronic stress-induced mitochondrial changes can causally contribute to the development of depression.

Amenability and geometric group theory

Ergodic and Geometric Group Theory (EGG)

See publications

©Yash Lodha

Group theory studies abstractions of symmetries of a physical object, or a theoretical space. For example, consider translations and rotations in the familiar three dimensional Euclidean space.

From a combinatorial perspective, a group is comprised of words in an alphabet that are equipped with the operation of concatenation and subject to a set of axioms. The combinatorial viewpoint has been dominant in twentieth century group theory, but in recent decades the perspective of viewing groups as geometric objects has been fruitful. As a consequence rich interactions have emerged between groups and geometry.

Two central themes within modern group theory are the study of groups through their actions on geometric objects and through their actions on measure spaces. My work combines both of these themes.

Many interesting examples of groups emerge as groups of homeomorphisms of the circle. Recently, new examples of such groups were introduced by Monod, and also by me in joint work with my PhD supervisor, Justin Moore. These examples provide new, elegant counterexamples to a question of the mathematician and physicist Jon von Neumann from 1929: Are there non amenable groups that do not contain non abelian free subgroups? This question, as well as the notion of amenability, emerged from von Neumann’s study of the Banach-Tarski paradox, one of the most striking phenomena in mathematics.

However, there are various examples of groups whose amenability remains a mystery. One of the goals of this project is to investigate the amenability of such groups. Another goal of this project is to understand examples of groups from various perspectives: fniteness properties, nonpositive curvature, normal subgroup structure, sofcity and various weaker notions of amenability.

(Start date of fellowship: 1 April 2015)

Cayley graph of a free group


Solar water splitting via perovskite photovoltaics and earth-abundant catalysts

Laboratory of Photonics and Interfaces (LPI)

See publications

©Jingshan Luo

Solar energy is plentiful enough to fulfill all of mankind’s energy demands as a green and renewable source. However, its diffuse and intermittent nature requires an efficient harvesting strategy and effective method for storage before large-scale utilization can be envisaged.

Inspired by natural photosynthesis, converting solar energy directly into chemical fuels through artificial photosynthesis is considered as one of the most promising ways to solve this challenge.

Among all the chemical fuels, hydrogen is the simplest form, and it can be generated through water splitting. Producing hydrogen fuel through water splitting with the energy from the sun has been considered as the Holy Grail of photoelectrochemistry. In order to be economically viable, the device needs to be efficient, cost effective and stable, which demands a system that combined of Earth-abundant materials and can be easily processed. Here we propose to combine the emergent perovskite solar cells with Earth-abundant water splitting catalysts for the highly efficient and low cost water splitting. This project will provide both scientific and economic impact.

(Start date of fellowship: 1 April 2015)

©Jingshan Luo and Alain Herzog
A: Perovskite solar cell B: NiFe layered double hydroxide water splitting catalyst

Theoretical foundations of high-dimensional statistical inference with non-linear models

Laboratory for Information and Inference Systems (LIONS)

See publications

©Luca Baldassare

Modern data set sizes are increasing at unprecedented rates, causing conventional methods for tasks such as data acquisition, inference, and learning to become impractical. Recently, there have been tremendous advances in these areas arising from the insight that despite the large data set sizes, the actual “information level” is typically much smaller. By designing data processing systems to operate at this level accordingly, significant savings of crucial resources are made possible.

These advances have been made possible by both new mathematical theory, and the development of practical methods to match the theory. However, there remain growing gaps between theory and practice due to the inaccuracy of commonly-used mathematical models in real-world settings. The goal of this project is to bridge these gaps by developing theoretical foundations using more accurate models that are specifically targeted at practical applications.

The key points of focus will include (i) handling non-linear models, which have an expressive power far beyond commonly-used linear models, finding applications in areas such as low-light imaging and computerized tomography; (ii) achieving robustness with respect to mismatched designs, where the assumed model differs from the true model.

(Start date of fellowship: 1 May 2015)

©Volkan Cevher
Dimensionality reduction: Since the length-N signal x has only a few non-zero entries, its relevant information is captured by a smaller number M of measurements in the signal y.

Aligned and region-selective deposition of organic nanowires for thin film transistor applications

Laboratory of Macromolecular and Organic Materials (LMOM)

See publications

©Shawn Koppenhoefer
Thomas Schmaltz

Large scale patterning of organic semiconductors for thin film transistor applications remains a challenge, due to the incompatibility of organic semiconductors with standard lithographical processes. Currently used methods to achieve patterned organic semiconductor films such as ink-jet printing or physical vapor deposition through stencil masks, are limited in terms of feature size, integration density and large-scale fabrication of electronic devices.

The proposed research project aims to develop a novel approach to address this problem, by separating the semiconductor growth from the patterning step. For this purpose, we intend to use organic semiconductor nanowires, prepared by anisotropic crystallization from solution. These one-dimensional systems will then be deposited on pre-patterned substrates containing a hydrophobic/hydrophilic contrast that is achieved by functionalizing the substrate with patterns of self-assembled monolayers.

With this approach, we aim for the region-selective deposition of the nanowires and even for an alignment, induced by the shape of the wettable patterns on the surface. As the deposition of the nanowires from dispersions occurs after the patterning step, their internal structure will not be affected by this process. The final goal of this project is to incorporate aligned and patterned nanowire networks into thin film transistor devices.

(Start date of fellowship: 1 May 2017)

©Thomas Schmaltz
a) Preparation of organic semiconductor nanowires from solution; b) deposition of these nanowires from dispersion of pre-patterned substrates that contain wettable and non-wettable regions; c) incorporation of patterned and aligned nanowire arrays in thin film transistor (TFT) devices.

Well-defined ultrathin microcapsules by carbonization of reactive molecular precursors in microfluidic systems

Laboratory of Macromolecular and Organic Materials (LMOM)

See publications

©Bjoern Schulte

Liposomes are formed from small amphiphiles and often serve as simplified model systems of biological membranes. The current project, which is also supported by the Alexander-von-Humboldt foundation, aims to prepare “carbosomes” as a novel type of vesicle-based material (Figure 1).

To this end, it combines the use of microfluidic systems, the self-assembly of reactive molecular carbon precursors at the oil-water interface, and their carbonization by UV irradiation at room temperature. Carbosomes are carbon microcapsules with a diameter on the order of micrometers, an ultrathin, extensively cross-linked carbon shell with a thickness of 2–4 nm, and a defined chemical surface functionalization.

Presumably, carbosomes combine the low shell thickness as well as the related permeability and possibility of directed transport of biological cells and synthetic liposomes with the mechanical rigidity and electrical properties of carbon nanomaterials. Thus, the project is expected to deliver insights into the mechanical stability and porosity of these microcapsules.

(Start date of fellowship: 1 November 2015)

©LMOM group
Figure 1. Schematic illustration of the approach for the preparation of carbon microcapsules. It combines a) the use of amphiphilic carbon precursors with b) microfluidic systems. c) After self-assembly at the oil-water interface subsequent irradiation with UV light leads to carbon microcapsules with an ultrathin shell.


New approach to airline and train passenger demand modelling using aggregated big data

Transport and Mobility Laboratory (TRANSP-OR)

See publications

©Shadi Sharif Azadeh

The use of mathematics and computer science to solve concrete operational problems is recognized as “operations research”. This field of research has been successful during the last decades to solve larger problems relevant to the industry and to the society especially in Transport. The availability of big data describing the behavior of the users of these systems allows to optimize the operations not only with respect to the running cost, but also maximizing the level of service provided by these systems. In this project, we aim at investigating new methodologies allowing to include a behavioral dimension in the optimization of transportation systems, by exploiting big data.

The bias resulting from inaccurate demand forecasts has a negative impact on the level of service that the company can offer. The objective of this research is to propose a data driven mathematical model. The aim is to capture passengers’ behavior while optimizing the network efficiency based on the service level. In addition, she aims at estimating substitution probabilities when resources change due to stock outs. The outcome is crucial for inventory control and resource allocation in transportation systems to keep the service level as high as possible while making the whole system profitable.

(Start date of fellowship: 1 May 2016)

©Shadi Sharif Azadeh
Choice based optimization models introduce a powerful tool to guarantee high quality service level to travellers in the field of transportation.

Multivalent cellulose nanocrystals as viral entry inhibitors (Crystal CAVE-IN)

See publications

©Cristina Pérez Vera

The emergence of global epidemics associated with viral pathogens, such as Human Immunodeficiency Virus (HIV) and influenza, calls for the development of novel strategies to treat and prevent viral infections at lower cost.

Polyanionic compounds and spherical nanoparticles have been extensively investigated as viral entry inhibitors, however rod-shaped cellulose nanocrystals (CNCs) have only recently been considered for nanomedicine applications. Cellulose nanocrystals offer a novel nanotechnological platform for unique libraries of viral entry inhibitors. By modifying CNC surfaces with compounds that mimic the human cell receptors which allow viral entry, viruses will be deceived and inactivated.

The goal of the present project is to develop a library of CNCs containing biomimetic surface functionalities targeted for HIV entry inhibition, followed by evaluation of their potential cytotoxicity and binding affinity to HIV envelope glycoproteins. Not only can these nanomaterials be produced utilizing low-cost chemistries, but also provide a renewable non-cytotoxic substrate that can be functionalized with a practically endless library of target-specific compounds. Overall, the Crystal CAVE-IN project hopes to open new approaches for nanomedicinal therapies and prevention for the global population suffering from disease caused by viral infections.

(Start date of fellowship: 1 March 2015)

©University of Canterbury, 1996
Production of biomimetic cellulose nanocrystals and hypothetical viral entry inhibition. Adapted from S. Eichhorn, Soft Matter, 2011,7, 303-315. Tree Artwork by Mark Harrington.