A5021681620- Computational Drug Discovery Methods
- Machine Learning in Materials Science
- Protein Structure and Dynamics
- Chemical Synthesis and Analysis
- Microbial Natural Products and Biosynthesis
- Cancer Treatment and Pharmacology
- Surface Chemistry and Catalysis
- Pharmacological Receptor Mechanisms and Effects
- Advanced Memory and Neural Computing
- Face and Expression Recognition
- Free Radicals and Antioxidants
- Enzyme Structure and Function
- Organic Chemistry Cycloaddition Reactions
- Tuberculosis Research and Epidemiology
- Enzyme Catalysis and Immobilization
- Plant biochemistry and biosynthesis
- Innovative Microfluidic and Catalytic Techniques Innovation
- Chemical Reaction Mechanisms
- Analytical Methods in Pharmaceuticals
- Various Chemistry Research Topics
- DNA and Nucleic Acid Chemistry
- Advanced Biosensing Techniques and Applications
- Chemical synthesis and alkaloids
- Cyclopropane Reaction Mechanisms
- Origins and Evolution of Life
Highland Community College - Illinois
2019-2025
Polish Academy of Sciences
2018-2024
Institute of Organic Chemistry
2018-2024
Wrocław University of Science and Technology
2012-2020
AGH University of Krakow
2017-2020
Training algorithms to computationally plan multistep organic syntheses has been a challenge for more than 50 years1-7. However, the field progressed greatly since development of early programs such as LHASA1,7, which reaction choices at each step were made by human operators. Multiple software platforms6,8-14 are now capable completely autonomous planning. But these 'think' only one time and have so far limited relatively simple targets, could arguably be designed chemists within minutes,...
The challenge of prebiotic chemistry is to trace the syntheses life's key building blocks from a handful primordial substrates. Here we report forward-synthesis algorithm that generates full network chemical reactions accessible these substrates under generally accepted conditions. This contains both reported and previously unidentified routes biotic targets, as well plausible abiotic molecules. It also exhibits three forms nontrivial emergence, molecules within can act catalysts downstream...
Applications of machine learning (ML) to synthetic chemistry rely on the assumption that large numbers literature-reported examples should enable construction accurate and predictive models chemical reactivity. This paper demonstrates abundance carefully curated literature data may be insufficient for this purpose. Using an example Suzuki–Miyaura coupling with heterocyclic building blocks─and a selected database >10,000 examples─we show ML cannot offer any meaningful predictions optimum...
General conditions for organic reactions are important but rare, and efforts to identify them usually consider only narrow regions of chemical space. Discovering more general reaction requires considering vast space derived from a large matrix substrates crossed with high-dimensional conditions, rendering exhaustive experimentation impractical. Here, we report simple closed-loop workflow that leverages data-guided down-selection, uncertainty-minimizing machine learning, robotic discover...
Machine learning can predict the major regio-, site-, and diastereoselective outcomes of Diels-Alder reactions better than standard quantum-mechanical methods with accuracies exceeding 90 % provided that i) diene/dienophile substrates are represented by "physical-organic" descriptors reflecting electronic steric characteristics their substituents ii) positions such relative to reaction core encoded ("vectorized") in an informative way.
The ability to estimate the acidity of C–H groups within organic molecules in non-aqueous solvents is important synthetic planning correctly predict which protons will be abstracted reactions such as alkylations, Michael additions, or aldol condensations. This Article describes use so-called graph convolutional neural networks (GCNNs) perform predictions on time scales milliseconds and with accuracy comparing favorably state-of-the-art solutions, including commercial ones. crux method train...
The prevalent assumption in computer-assisted synthesis planning has been to rely on the wealth of reaction data and consideration this vast knowledge base at every stage route planning. Yet even if equipped with all requisite individual transforms state-of-the-art search algorithms, existing programs struggle when confronted advanced targets, such as complex peptides work considers. By contrast, searches are constrained by hierarchical logic, dictating which subsets reactions apply...
Generation of structural analogs to “parent” molecule(s) interest remains one the important elements drug development. Ideally, such should be synthesizable by concise and robust synthetic routes....
This work describes a method to vectorize and Machine-Learn, ML, non-covalent interactions responsible for scaffold-directed reactions important in synthetic chemistry. Models trained on this representation predict correct face of approach ca. 90 % Michael additions or Diels-Alder cycloadditions. These accuracies are significantly higher than those based traditional ML descriptors, energetic calculations, intuition experienced chemists. Our results also emphasize the importance models being...
Abstract Machine learning can predict the major regio‐, site‐, and diastereoselective outcomes of Diels–Alder reactions better than standard quantum‐mechanical methods with accuracies exceeding 90 % provided that i) diene/dienophile substrates are represented by “physical‐organic” descriptors reflecting electronic steric characteristics their substituents ii) positions such relative to reaction core encoded (“vectorized”) in an informative way.
When an organometallic catalyst is tethered onto a nanoparticle and embedded in monolayer of longer ligands terminated "gating" end-groups, these groups can control the access orientation incoming substrates. In this way, nonspecific become enzyme-like: it select only certain substrates from substrate mixtures and, quite remarkably, also preorganize such that some their otherwise equivalent sites react. For simple, copper-based click reaction for gating charged groups, both substrate-...
It is known that when catalytic nanoparticles are functionalized with charged ligands, the polarity of these ligands can selectively control approach either (+) or (−) substrates, effectively rendering particles' activity charge-selective. In such experiments, however, role counterions surrounding generally not considered. The present work demonstrates counterions—despite being only loosely bound—can have a dramatic effect on on-particle catalysis. particular, same but different sizes,...
A computer program for retrosynthetic planning helps develop multiple "synthetic contingency" plans hydroxychloroquine and also routes leading to remdesivir, both promising but yet unproven medications against COVID-19. These are designed navigate, as much possible, around known patented commence from inexpensive diverse starting materials, so ensure supply in case of anticipated market shortages commonly used substrates. Looking beyond the current COVID-19 pandemic, development similar...
The concept of the polarization justified Fukui functions has been tested for set model molecules: imidazole, oxazole, and thiazole. Calculations have based on molecular polarizability analysis, which makes them a potentially more sensitive analytical tool as compared to classical density functional theory proposals, typically built electron only. Three selected molecules show distinct differences in their reactivity patterns, despite very close geometry electronic structure. maps plane...
Currently developed protocols of theozyme design still lead to biocatalysts with much lower catalytic activity than enzymes existing in nature, and, so far, the only avenue improvement was vitro laboratory-directed evolution (LDE) experiments. In this paper, we propose a different strategy based on "reversed" methodology mutation prediction. Instead common "top-down" approach, requiring numerous assumptions and vast computational effort, argue for "bottom-up" approach that is fields derived...
Artificial Intelligence algorithms are used to identify “progeny” drugs that similar the “parents” already being tested against COVID-19. These assess similarity not only by molecular make-up of molecules, but also “context” in which specific functional groups arrangedand/or three-dimensional distribution pharmacophores. The parent-progeny relationships span same-indication (mostly antivirals) as well those “progenies” have different and perhaps less intuitive primary indications (e.g.,...
At distances shorter than equilibrium, electrostatic interactions seem to be a more robust indicator of relative molecular dimer stability accurate electronic structure approaches. We arrive at this conclusion by investigating the nonparametric correlation between reference interaction energies equilibrium geometries (coupled cluster with singles, doubles, and perturbative triples complete basis set limit, ΔECCSD(T)CBS,ref) its various approximate values obtained range for training 22...
Abstract This work describes a method to vectorize and Machine‐Learn, ML, non‐covalent interactions responsible for scaffold‐directed reactions important in synthetic chemistry. Models trained on this representation predict correct face of approach ca. 90 % Michael additions or Diels–Alder cycloadditions. These accuracies are significantly higher than those based traditional ML descriptors, energetic calculations, intuition experienced chemists. Our results also emphasize the importance...
Fatty acid amide hydrolase (FAAH) is an enzyme responsible for the deactivating hydrolysis of fatty ethanolamide neuromodulators. FAAH inhibitors have gained considerable interest due to their possible application in treatment anxiety, inflammation, and pain. In context inhibitor design, availability reliable computational tools predicting binding affinity still a challenging task, it now well understood that empirical scoring functions several limitations principle could be overcome by...
We propose a simple atomic multipole electrostatic model to rapidly evaluate the effects of mutation on enzyme activity and test its performance wild-type mutant ketosteroid isomerase. The predictions our are similar those obtained with symmetry-adapted perturbation theory at fraction computational cost. further show that this approach is relatively insensitive precise amino acid side chain conformation in mutants may thus be useful (re)design.
Artificial Intelligence algorithms are used to identify “progeny” drugs that similar the “parents” already being tested against COVID-19. These assess similarity not only by molecular make-up of molecules, but also “context” in which specific functional groups arrangedand/or three-dimensional distribution pharmacophores. The parent-progeny relationships span same-indication (mostly antivirals) as well those “progenies” have different and perhaps less intuitive primary indications (e.g.,...