ERC FUNDED PROJECTS

There has been a long-standing quest to observe chemical reactions at low temperatures where reaction rates and pathways are governed by quantum mechanical effects. So far this field of Quantum Chemistry has been dominated by theory. The difficulty has been to realize in the laboratory low enough collisional velocities between neutral reactants, such that the quantum wave nature could be observed. Recently we have demonstrated a new way of studying cold reactive collisions by magnetically merging two fast neutral supersonic beams. After 40 years where the reactive scattering temperature was limited to above 5 K we were able to continuously tune collision energies from hundreds of Kelvin down to 10 mK temperature, a reduction of almost three orders of magnitude [A. B. Henson et. al, Science 338, 234, 2012]. Importantly, we were able to show that at low temperatures quantum effects start dominating reactive dynamics with the first observation of orbiting resonances in a reactive collision. We propose to extend our novel method to study chemical reactions in the regime of Cold Chemistry where the reactants’s de Broglie wavelength becomes larger compared to the characteristic interaction range. Theoretical predictions at low temperatures are extremely sensitive to the parameters used, routinely differing by orders of magnitude leading to contradictions waiting to be settled by experiment. Our ability to reach low enough collision energies and resolve scattering resonances will be used to bring a radical change to transient species spectroscopy. We believe that our work will not only test the central tenets of Quantum Chemistry, but will also provide valuable information to other fields, such as Astrochemistry helping to understand the synthesis of various molecules in interstellar space at temperatures 10 K and below.

1 982 908 €

Start date: 2014-01-01, End date: 2018-12-31

The aim of my project is to establish a multidisciplinary platform for quantitative biophysical analysis of protein-protein interactions in health and disease as a basis for drug design: (1) Analyzing protein-protein interactions at the molecular level in healthy systems; (2) Understanding what goes wrong in disease at the molecular level; (3) Development of drugs that will restore the biological system to its healthy conditions. My team will apply this approach to establish the concept of shifting the oligomerization equilibrium of proteins as a therapeutic strategy. I will expand the concepts of allosteric inhibitors and chemical chaperones, and develop the “shiftides”: peptides that shift the oligomerization equilibrium of a protein to modulate its activity, as a new and widely applicable methodology for drug design. I will apply this concept for: (1) inhibiting a protein by binding preferentially to the inactive oligomeric state and shifting the oligomerization equilibrium of the protein towards it; I have demonstrated the feasibility of this approach and developed promising anti-HIV peptides that inhibit the HIV-1 integrase and consequently HIV-1 replication in cells by shifting the integrase oligomerization equilibrium from the active dimer to the inactive tetramer. My team will further develop these peptides, and apply the same approach to inhibit the HIV proteins reverse transcriptase and protease; (2) Activating a protein by binding preferentially to the active oligomeric state and shifting the oligomerization equilibrium towards it: This will be applied for activation of the tumor suppressor p53, by shifting its oligomerization equilibrium from the inactive dimer to the active tetramer. Such shiftides will serve as anti-cancer lead compounds. My project will open new doors in the field of drug design, and at the end of the five-year period will result in a general new methodology to affect protein function for medical purposes.

1 250 000 €

Start date: 2008-07-01, End date: 2013-06-30

Critical for the function of many proteins, allosteric communication involves transmission of the effect of binding at one site of a protein to another through conformational changes. Yet the structural and dynamic basis for allostery remains poorly understood. In particular, there is no method to follow coordinated large-scale motions of domains and subunits in proteins as they occur. Since the subunits of allosteric proteins often contain multiple domains, any such method entails probing the dynamics along several intra-protein distances simultaneously. This proposal aims at ameliorating this deficiency by creating the experimental framework for exploring time-dependent coordination of allosteric transitions of multiple units within proteins. Our methodology will rely on single-molecule FRET spectroscopy with multiple labels on the same protein and advanced analysis. We will explore fundamental issues in protein dynamics: relative motions of domains within subunits, propagation of conformational change between subunits, and synchronization of these motions by effector molecules. To investigate these issues, we have carefully selected three model systems, each representing an important scenario of allosteric regulation. While the homo-oligomeric protein-folder GroEL conserves symmetry in a concerted transition between major structural states, the symmetry of the homo-oligomeric disaggregating machine ClpB is broken via a sequential transition. Symmetry is attained only after binding to DNA and ligands in the third system, the family of RXR heterodimers. This exciting project will provide the very first catalogue of coordinated and time-ordered motions within and between subunits of allosteric proteins and the first measurement of the time scale of the conformational spread through a large protein. It will enhance dramatically our understanding of how allostery contributes to protein function, influencing future efforts to design drugs for allosteric proteins.

2 484 722 €

Start date: 2017-05-01, End date: 2022-04-30

Lately, deep learning (DL) has become one of the most powerful machine learning tools with ground-breaking results in computer vision, signal & image processing, language processing, and many other domains. However, one of its main deficiencies is the lack of theoretical foundation. While some theory has been developed, it is widely agreed that DL is not well-understood yet. A proper understanding of the learning mechanism and architecture is very likely to broaden the great success to new fields and applications. In particular, it has the promise of improving DL performance in the unsupervised regime and on regression tasks, where it is currently lagging behind its otherwise spectacular success demonstrated in massively-supervised classification problems. A somewhat related and popular data model is based on sparse-representations. It led to cutting-edge methods in various fields such as medical imaging, computer vision and signal & image processing. Its success can be largely attributed to its well-established theoretical foundation, which boosted the development of its various ramifications. Recent work suggests a close relationship between this model and DL, although this bridge is not fully clear nor developed. This project revolves around the use of sparsity with DL. It aims at bridging the fundamental gap in the theory of DL using tools applied in sparsity, highlighting the role of structure in data as the foundation for elucidating the success of DL. It also aims at using efficient DL methods to improve the solution of problems using sparse models. Moreover, this project pursues a unified theoretical framework merging sparsity with DL, in particular migrating powerful unsupervised learning concepts from the realm of sparsity to that of DL. A successful marriage between the two fields has a great potential impact of giving rise to a new generation of learning methods and architectures and bringing DL to unprecedented new summits in novel domains and tasks.

1 499 375 €

Start date: 2017-10-01, End date: 2022-09-30