Unconventional Catalysis in Organic Chemistry: a Computational Mechanistic Study

Catalysis plays a vital role in modern synthetic chemistry. However, even if conventional catalysis (organo-catalysis, metal-catalysis and enzyme-catalysis) has provided outstanding results, various unconventional ways to make chemical reactions more effective appear now very promising. Computational methods can be of great help to reach a deeper comprehension of these chemical processes. The methodologies employed in this thesis are Quantum-Mechanical (QM), Molecular Mechanics (MM) and hybrid Quantum-Mechanical/Molecular Mechanics (QM/MM) methods. In this abstract the results are briefly summarised. The first unconventional catalysis investigated consists in the application of Oriented External Electric Fields (OEEFs) to SN2 and 4e-electrocyclic reactions. SN2 reactions with back-side mechanism can be catalysed or inhibited by the presence of an OEEF. Moreover, OEEFs can inhibit back-side mechanism (Walden inversion of configuration) and promote the naturally unfavoured front-side mechanism (retention of configuration). Electrocyclic ring opening reaction of 3-substituted cyclobutene molecules can occur with inward or outward mechanisms depending on the nature of substituent groups on the cyclobutene structure (torquoselectivity principle). OEEFs can catalyse the naturally favoured pathway or circumvent the torquoselectivity principle leading to different stereoisomers. The second case study is based on Carbon Nanotubes (CNTs) working as nano-reactors: the reaction of ethyl chloride with chloride anion inside CNTs was investigated. In addition to the SN2 mechanism, syn and anti-E2 reactions are possible. These reactions inside CNTs of different radii were examined with hybrid QM/MM methods, finding that these processes can be both catalysed and inhibited by the CNT diameter. The results suggest that electrostatic effects govern the activation energy variations inside CNTs. Finally, a new biochemical approach, based on the use of DNA catalyst was investigated at QM level. Deoxyribozyme 9DB1 catalyses the RNA ligation allowing the regioselective formation of the 3'-5' bond, following an addition-elimination two-step mechanism.

Supporting Scientific Research Through Machine and Deep Learning: Fluorescence Microscopy and Operational Intelligence Use Cases

Although the debate of what data science is has a long history and has not reached a complete consensus yet, Data Science can be summarized as the process of learning from data. Guided by the above vision, this thesis presents two independent data science projects developed in the scope of multidisciplinary applied research. The first part analyzes fluorescence microscopy images typically produced in life science experiments, where the objective is to count how many marked neuronal cells are present in each image. Aiming to automate the task for supporting research in the area, we propose a neural network architecture tuned specifically for this use case, cell ResUnet (c-ResUnet), and discuss the impact of alternative training strategies in overcoming particular challenges of our data. The approach provides good results in terms of both detection and counting, showing performance comparable to the interpretation of human operators. As a meaningful addition, we release the pre-trained model and the Fluorescent Neuronal Cells dataset collecting pixel-level annotations of where neuronal cells are located. In this way, we hope to help future research in the area and foster innovative methodologies for tackling similar problems. The second part deals with the problem of distributed data management in the context of LHC experiments, with a focus on supporting ATLAS operations concerning data transfer failures. In particular, we analyze error messages produced by failed transfers and propose a Machine Learning pipeline that leverages the word2vec language model and K-means clustering. This provides groups of similar errors that are presented to human operators as suggestions of potential issues to investigate. The approach is demonstrated on one full day of data, showing promising ability in understanding the message content and providing meaningful groupings, in line with previously reported incidents by human operators.