Fellows from 3rd call

Abstract coming soon! 

TBC

Abstract coming soon! 

TBC

Reducing food waste in industrial processes could be a significant step towards a sustainable society. Many of the machineries currently in use were designed with empirical methods, which science can drive in the right direction. For example, out of 1 kg of raw rice, the best-in-class processes today result in around 550g of white rice and 150g of broken rice, which is a lower value product. Considering the 500 million tons processed annually, a reduction of a few percent in brokens, or an increase of a few percent in energy efficiency could have a significant ecologic and economic impact. This project aims to use a combination of simulations, modelling and experiments to understand the interplay between granular media representing real crops and its confining media. The goal is also to identify which parameters are critical for the jamming transition, how it affects particle wear and, ultimately, how to control it.

The diminishment of motor ability is one of the most disabling aspects of ageing. Using tractable models for motor ageing, we have found a key role for an understudied form of synaptic communication, miniature neurotransmission, in maintaining motor ability during ageing. In addition, the levels of Trio, an evolutionarily conserved Guanine nucleotide Exchange Factor (GEF) (Banerjee et al., 2021), declined at NMJ synapses with age. Remarkably, we have also shown that motor synapses undergo bouton fragmentation when neurotransmission was inhibited by blocking postsynaptic muscle neurotransmitter receptors, suggesting an unidentified retrograde signal required to maintain structures. These results further suggest that loss of presynaptic Trio or this unknown retrograde signal is essential to maintain the capacity of synapses to sustain the high intensities of neurotransmitter release necessary to maintain robust motor function. This study will build on these results to identify key molecules that can ameliorate motor synapse age-dependent deterioration and preserve motor ability in aged animals. Achievement of these goals could suggest novel therapeutic strategies to ameliorate the universal problem of the decline of motor ability in humans as we age.

Developing robotic approaches to harvesting would relieve the pressure on agriculture workers, and improve sustainability and food security. Furthermore, robots offer more controllability of agriculture operations, offering precision harvesting which could reduce waste and improving the quality of the produce generated. Although developments in robotic hardware and machine learning are making robotic harvesting possible, they do not utilize tactile sensors, which limits their use to more fragile or delicate crops. We propose exploring how developing soft sensorized and human-like robotic hand can be used to allow harvesting of a range of soft produce. By leveraging tactile, visual information and imitation learning algorithms based on human demonstrations, the proposed robotic hand can perform dexterous and robust manipulation that can improve the quality of the produce and the efficiency of harvesting.

Constraining neuronal activity within physiological ranges is key for reliable brain function. In parallel, there must also exist the flexibility necessary for learning and memory formation during an organism’s lifetime. This study aims to create impact in healthcare, by providing a detailed molecular understanding of synaptic scaling, an essential adaptation response to changing intrinsic or extrinsic factors in the brain. This will offer a strong foundation to investigate the extent to which aberrant homeostatic signalling could contribute to the disease pathogenesis of a plethora of neuropsychiatric conditions. As an added contribution to sustainability, we have evidence to show that synaptic scaling can be directly influenced by environmental changes. Hence, we aim to establish and detail the first direct link between mechanisms of plasticity and the behavioural adaption of an organism. In the broader view, this could help understand the neuronal effects upon animals and humans of a rapidly warming globe.

Within the transition towards a more inclusive and sustainable society, new solutions are arising in the field of assistive robotics. Lower-limbs exoskeletons have shown promising results to improve walking and balance, with the potential to increase the mobility of the elderly population and therefore promote a more inclusive and accessible community. Despite numerous existing devices in the literature, researchers are still investigating ways to achieve a safe and compliant collaboration between humans and wearable robots. To tackle this challenge, my thesis is focused on the user-centred design of assistive strategies for a powered hip exoskeleton. The control strategy will address the challenge of assisting diversified activities of daily living and, after laboratory assessments, it will be tested in outdoor environments. Qualitative and quantitative measurements will evaluate the effectiveness of the exoskeleton assistance in decreasing the perceived walking effort, avoiding sedentary lifestyle and reducing the risk of balance loss. 

Producing high-value chemicals from inexpensive renewable sources could help to develop sustainable industrial processes. Biomass offers a cheap and abundant source of carbon building blocks. However, the adoption of these renewable feedstocks is limited by the lack of routes to important chemicals starting from these building blocks. My project aims to systematically explore the possible routes leading to biomass-based high-value substances. Thanks to AI and data-driven techniques, I’m generating and screening thousands of possible routes in the biomass chemical space, scoring the most promising transformations. These selected routes can be experimentally verified to discover more sustainable ways to produce valuable molecules. 

Designing proteins that can dynamically and reversibly switch between different states remains challenging. To address this, we developed a computational framework for engineering light-mediated switchable properties in target molecules. We leveraged the LOV2 optogenetic domain to control the conformational dynamics of de novo designed scaffolds, creating light-responsive protein switches. We anticipate that functionalizing these switches will lead to impactful biological applications. One important goal of this project is to enhance and diversify optogenetic approaches for restoring vision in patients suffering from Retinal Degenerative Diseases (RDDs) through the use of designed photo-switchable proteins, ultimately improving their quality of life. To achieve this goal, we aim to engineer a novel class of light-inducible G-protein activators that expose a soluble peptide designed to activate selective G-proteins upon light exposure

The goal of personalised nutrition is to preserve or increase health using genetic, phenotypic, medical, nutritional, or other relevant information about an individual to deliver tailored recommendations. The rational being: a) response to dietary factors varies between individuals; b) personalised diets have more impact than general diets; c) personalisation supports long-term behaviour change. With the help of machine learning models trained on a large-scale digital nutritional tracking cohort called Food&You (>1000 participants), our aim is to control glucose responses in the general populace by recommending personalized healthier food/diet options, which could in principle, also be linked to sustainable food choices.

Understanding how neurons in the brain function together (and become dysfunctional in disease) is essential to develop novel treatments for neurological or psychiatric disorders. For decades, motor neuron research has primarily focused on excitatory motor projections — when motor neurons fire, we get an excitatory response in the form of muscle movement. However, inhibitory projections might also contribute to motor control, allowing the correct motor neuron to fire at the correct time. This project aims to use a tractable model system and experimental data to characterise inhibitory projections in motor circuits and understand how their dysfunction contributes to neurological disease. 

The rise of drug-resistant bacteria, specifically those producing β-lactamase enzymes that can degrade antibiotics, poses a serious global health threat. Of particular concern is their increasing resistance to carbapenem antibiotics, which are last-resort drugs used in clinics for treating bacterial infections. These β-lactamase enzymes can also operate outside bacterial cells by being transported in small membrane structures called outer membrane vesicles (OMVs), which contain various proteins. When Gram-negative bacteria face stressful conditions, they release these OMVs, extending the range of β-lactamase activity and shielding bacterial populations from antibiotics, including those that would be otherwise vulnerable to them. Our research aims to uncover how these enzymes associate with OMVs. We will apply molecular dynamics simulations and biochemistry experiments to determine which are the fundamental physicochemical properties of these enzymes that make them more or less likely to be found in OMVs, also uncovering potential druggable targets in these molecules.

To cover Swiss energy demands according to the Swiss Energy Strategy 2050, wind and solar energy production must be greatly increased. A large part of this energy can be extracted via wind turbines in the Alps as winds are strong and occur often. Due to the complex terrain of mountains and slopes, the wind is hard to predict which hinders the progress of wind energy in Switzerland. Mathematical models exist but lack quality wind data at the height of the wind turbine to validate and improve wind energy prediction. A wind-doppler LiDAR uses lasers which reflect on aerosols such as water vapour and dust which provides us with information about the wind speed around the turbine. However, such wind energy measurement campaigns in the Alps are challenging and require innovative solutions. By using machine learning algorithms, we are able to connect this to the weather to fine-tune the models.

The proposed project focuses on studying the territorial governance and planning viewed in the context of crisis. Based on the empirical setting of urbanisation in Ukraine, the project will specifically investigate the effects of territorial regulations associated with “decentralisation”, expressed in the decision-making and planning process at a local level and introduced in the context of multiple crises, linked to the ongoing war, on territorial resilience.
The investigation derives from two case studies of territories characterised by the recent administrative merger of core cities with their productive hinterland and embedded into diverging emergency and long-term planning scenarios linked to the war-related crises, thus allowing to situate and review the interaction between decentralisation-oriented and risk-induced discourses in planning. The study will provide a more profound understanding of the potency of territorial resilience applied to territorial governance and spatial planning based on a qualitative understanding of territory and the emergence of regulations that govern land use and production of the built environment within it. It will also contribute to a general debate on the increased agency of municipalities and metropolitan areas in generating territorial resilience.

Diamond grown by chemical vapor deposition (CVD) is drawing increased attention from both scientific community and industry as a highly promising material for photonics and quantum sensing. The objective of this project is to advance the performances of CVD grown diamond thin films and large scale wafers in terms of fundamental physical properties, such as thermal conductivity or optical transparency and spin coherence time for photonics and quantum sensing. We aim to grow and micro-structure high-quality single crystalline diamond (thin layers) and highly oriented polycrystalline diamond (on large area substrates) for applications in photonic integrated circuits and quantum sensors.

My PhD project will make a contribution to the Energy-Food-Water-Land use connections. In this context packaging materials have an important role to play, due to their quantites, their energy content, their properties to conserve and protect goods, including food, and enventually also because of their potential to cause environmental problems at the End-of-Life (EoL) with potential pollution of water and soil bodies. In GR-LUD at EPFL, research aims to close the materials cycles or to recovert energy, in cases where materials re-use, recycling or recovery fail to reduce the overall environmental impacts. During my PhD I will study chemical-physical properties with a unique equipment available at EPFL aiming to predict the biodegradability of packaging materials to support the designing of EoL treatment processes for materials and energy recovery.