Article
Biochemical Research Methods
Vyacheslav Tretyachenko, Vaclav Voracek, Radko Soucek, Kosuke Fujishima, Klara Hlouchova
Summary: CoLiDe is a computational tool that offers precise control over protein sequence composition, length, and diversity for designing combinatorial libraries. It uses an evolutionary approach to provide solutions for degenerate DNA templates and can prepare purified protein library samples with up to 10^11 - 10^12 unique sequences. CoLiDe presents a composition-centric approach to protein design for different functional phenomena.
Article
Biotechnology & Applied Microbiology
Francesca V. Gambacorta, Joshua J. Dietrich, Justin J. Baerwald, Stephanie J. Brown, Yun Su, Brian F. Pfleger
Summary: This study aimed to identify an isobutanol pathway cassette that can support the growth of a non-ethanol producing S. cerevisiae. By screening a combinatorial library of pathway enzymes, the researchers isolated a cassette that was able to produce 364 mg/L isobutanol under aerobic conditions. This approach shows promise for rewiring the metabolism of S. cerevisiae to produce other fermentation products.
FRONTIERS IN BIOENGINEERING AND BIOTECHNOLOGY
(2022)
Article
Chemistry, Medicinal
Yuliana Zabolotna, Dmitriy M. Volochnyuk, Sergey Ryabukhin, Kostiantyn Gavrylenko, Dragos Horvath, Olga Klimchuk, Oleksandr Oksiuta, Gilles Marcou, Alexandre Varnek
Summary: Most existing computational tools for de novo library design focus on generating, selecting, and combining structural motifs to form new library members. However, these approaches appear to be more theoretical and disconnected from reality due to the lack of a direct link between the chemical space of the retrosynthesized fragments and the pool of available reagents. This paper presents a new open-source toolkit called Synthons Interpreter (SynthI), which merges these two chemical spaces into a single synthons space.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2022)
Article
Green & Sustainable Science & Technology
Mauro D'Apuzzo, Azzurra Evangelisti, Daniela Santilli, Sofia Nardoianni, Giuseppe Cappelli, Vittorio Nicolosi
Summary: With the increasing emphasis on new soft mobility in urban areas, it is becoming more important to provide effective speed control measures for vehicular traffic. Vertical traffic calming measures, which are based on vehicle vertical deflection, have historically been the most widespread. However, the different speed behaviors of different vehicles on specific road profiles have been evidenced, leading to the need for investigating the dynamics underlying this phenomenon and developing a new approach to the design of vertical traffic calming devices.
Article
Chemistry, Multidisciplinary
Gael Schaeffer, Marcel J. Eleveld, Jim Ottele, Peter C. Kroon, Pim W. J. M. Frederix, Shuo Yang, Sijbren Otto
Summary: The challenge of understanding how chemistry leads to biology is highlighted in contemporary science. This paper focuses on achieving life-like properties in synthetic chemical systems, which are usually deterministic. However, natural phenomena are often stochastic in nature. The authors report on the emergence of two different self-replicators from a mixture of synthetic molecules in a stochastic fashion, leading to variations in their ratios. The variation is attributed to a stochastic nucleation process, which is more pronounced near a phase boundary. This stochasticity is unique in synthetic self-replicators systems.
JOURNAL OF THE AMERICAN CHEMICAL SOCIETY
(2022)
Article
Chemistry, Physical
A. Usha Vijayakumar, N. Aloni, V. Thazhe Veettil, G. Rahamim, S. S. Hardisty, M. Zysler, S. Tirosh, D. Zitoun
Summary: This study systematically investigates the performance of a ternary Ni-Fe-Co oxide gradient library for the oxygen evolution reaction (OER). The combinatorial approach used in this study allows for a faster and more reliable investigation compared to the traditional step by step approach. The results provide a foundation for exploring other mixed-metal oxide combinations.
ACS APPLIED ENERGY MATERIALS
(2022)
Article
Multidisciplinary Sciences
Jaime Nebot, Juan A. Pena, Carmelo Lopez Gomez
Summary: This paper presents a new methodology of product shape definition, utilizing 3D morphing techniques and genetic algorithms to create unique and innovative product designs. The approach allows designers to detach themselves during the design process, providing unexpected geometric solutions and overcoming limitations of traditional 3D modeling programs.
Article
Chemistry, Multidisciplinary
Gonna Somu Naidu, Seok-Beom Yong, Srinivas Ramishetti, Riccardo Rampado, Preeti Sharma, Assaf Ezra, Meir Goldsmith, Inbal Hazan-Halevy, Sushmita Chatterjee, Anjaiah Aitha, Dan Peer
Summary: Ionizable lipid-based nanoparticles (LNPs) were synthesized by modifying hydrophobic tail chains and linkers, and stable LNPs were formed with the help of other lipids and a microfluidic mixing approach. Using various mRNA and animal models, specific lipids for efficient cell-type specific mRNA delivery were identified. In vitro assays showed that a combination of branched ester tail chains and hydroxylamine linker negatively affects mRNA delivery efficiency. In vivo studies identified superior liver-trophic and cell type-specific lipid nanoparticles. Comparisons with SM-102-based LNPs showed cell-specific mRNA delivery efficiency and toxicity. This study suggests that tail and linker combination can drive novel functionality of LNPs in vivo.
Article
Multidisciplinary Sciences
Azalia Mirhoseini, Anna Goldie, Mustafa Yazgan, Joe Wenjie Jiang, Ebrahim Songhori, Shen Wang, Young-Joon Lee, Eric Johnson, Omkar Pathak, Azade Nazi, Jiwoo Pak, Andy Tong, Kavya Srinivasa, William Hang, Emre Tuncer, Quoc V. Le, James Laudon, Richard Ho, Roger Carpenter, Jeff Dean
Summary: Machine learning tools are employed to accelerate chip layout design by treating it as a reinforcement learning problem and utilizing neural networks to generate high-performance layouts. The deep reinforcement learning approach presented in the study can automatically produce chip floorplans in under six hours that are superior or comparable to those created by humans in key metrics like power consumption, performance, and chip area.
Article
Chemistry, Medicinal
Thomas Liphardt, Thomas Sander
Summary: We propose an efficient algorithm for substructure search in combinatorial libraries defined by synthons. Our method improves on existing approaches by introducing powerful heuristics and fast fingerprint screening. With this, we achieve quick response times for searches in large combinatorial libraries. We have also made the Java source code available under the BSD license and implemented tools for substructure search in custom combinatorial libraries.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2023)
Article
Green & Sustainable Science & Technology
Karolina Dukala, Joanna Pyrkosz-Pacyna, Rafal Czarny
Summary: Design thinking (DT) is a popular method of problem-solving and idea generation in creative teams. However, there is limited evidence for its effectiveness and a lack of coherence in its underlying theory and definition. To address these issues, a new approach called the design thinking method (DTMethod) was introduced and tested among teams working on a specific task. The study found that teams using the DTMethod achieved more favorable results in terms of utility and meeting requirements, but not in terms of cost or time efficiency. Individuals using the DTMethod reported experiencing less positive emotions but were overall more satisfied with the experience and rated teamwork more favorably in terms of cohesiveness and team effort.
Article
Computer Science, Information Systems
Asif I. Omi, Zeba N. Zafar, Hussain Al-Shakhori, Aubrey N. Savage, Rakibul Islam, Mohammad A. Maktoomi, Christine Zakzewski, Praveen Sekhar
Summary: This paper presents a new analytical design technique for a three-section wideband Wilkinson power divider, utilizing the dual-frequency behavior of commensurate transmission lines for analysis and introducing rigorous design equations for odd-mode analysis. Measurement of a fabricated prototype using this technique demonstrates excellent performance exceeding the minimum requirements for return loss, insertion loss, and isolation over a wide bandwidth.
Article
Energy & Fuels
Linyuan Wen, Bozhou Wang, Tao Yu, Weipeng Lai, Jinwen Shi, Maochang Liu, Yingzhe Liu
Summary: An advanced approach combining combinatorial library design and high-throughput screening was developed to accelerate the search for CHONF-containing energetic materials with higher energy density. Approximately two thousand target molecules with promising properties were discovered, and the effectiveness of the approach was verified through DFT calculations. The study aims to provide new insights into the fast discovery of novel high-energy-density materials.
Article
Computer Science, Artificial Intelligence
Hien Ngoc Nguyen, Ganix Lasa, Ion Iriarte, Ariane Atxa, Gorka Unamuno, Gurutz Galfarsoro
Summary: The study introduces a multidimensional design methodology called DIMAND, which incorporates key design elements such as life-cycle service design, stakeholder networks, new service development methods, and design skills to enhance the effectiveness of advanced service design.
ADVANCED ENGINEERING INFORMATICS
(2022)
Article
Chemistry, Medicinal
Marko Jukic, Sebastjan Kralj, Anja Kolaric, Urban Bren
Summary: Peptides, as active ingredients of drugs and important tools in medical research, pose a challenge for computational methods. We propose an in silico workflow using CmDock to generate and prioritize peptide libraries, and successfully identified tetrapeptide ligands that bind to antibody Fc regions. Our results align with existing scientific literature and we suggest a developing in silico library design workflow to overcome the combinatorial problem of in vitro peptide libraries.
Article
Chemistry, Medicinal
Vendy Fialkova, Jiaxi Zhao, Kostas Papadopoulos, Ola Engkvist, Esben Jannik Bjerrum, Thierry Kogej, Atanas Patronov
Summary: The researchers have proposed a novel drug design tool called LibINVENT, capable of rapidly designing chemical libraries with shared cores and maximizing desirable properties. In addition, users can list specific chemical reactions for library creation, making this tool flexible for lead optimization.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2022)
Article
Chemistry, Medicinal
Rocio Mercado, Esben J. Bjerrum, Ola Engkvist
Summary: The impact of different graph traversal algorithms on molecular graph generation was explored in this study. Using a breadth-first traversal led to better coverage of training data features compared to a depth-first traversal, but overtraining can make the results with either graph traversal algorithm identical.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2022)
Article
Infectious Diseases
Wilson Lim, Bertrand Nyuykonge, Kimberly Eadie, Mickey Konings, Juli Smeets, Ahmed Fahal, Alexandro Bonifaz, Matthew Todd, Benjamin Perry, Kirandeep Samby, Jeremy Burrows, Annelies Verbon, Wendy van de Sande
Summary: Eumycetoma is a chronic subcutaneous neglected tropical disease caused by various fungal agents. Current antifungal drugs have low cure rates, and there is a need for novel treatments. In this study, four compounds (fenbendazole, carbendazim, tafenoquine, and MMV1578570) were found to inhibit the growth of five species of fungi that cause eumycetoma. These compounds showed promising results in in vivo experiments as well.
PLOS NEGLECTED TROPICAL DISEASES
(2022)
Editorial Material
Infectious Diseases
Claudia Daubenberger, Jeremy N. Burrows
LANCET INFECTIOUS DISEASES
(2022)
Article
Virology
Lucca R. Policastro, Isabela Dolci, Andre S. Godoy, Jose V. J. Silva Junior, Uriel E. A. Ruiz, Igor A. Santos, Ana C. G. Jardim, Kirandeep Samby, Jeremy N. Burrows, Timothy N. C. Wells, Laura H. V. G. Gil, Glaucius Oliva, Rafaela S. Fernandes
Summary: Chikungunya virus is the causative agent of chikungunya fever, a specific drug against this virus is needed for treatment.
Article
Chemistry, Medicinal
Sara Romeo Atance, Juan Viguera Diez, Ola Engkvist, Simon Olsson, Rocio Mercado
Summary: This research proposes a new reinforcement learning scheme to fine-tune graph-based deep generative models for de novo molecular design. The proposed approach can successfully guide a pretrained generative model to generate molecules with specific properties, even if these molecules are not present in the training set.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2022)
Article
Biology
Arwa Raies, Ewa Tulodziecka, James Stainer, Lawrence Middleton, Ryan S. Dhindsa, Pamela Hill, Ola Engkvist, Andrew R. Harper, Slave Petrovski, Dimitrios Vitsios
Summary: In this study, a stochastic semi-supervised ML framework called DrugnomeAI was developed to estimate the druggability likelihood for every protein-coding gene in the human exome. The tool generates exome-wide predictions based on known drug targets and provides specialized models stratified by disease type or drug therapeutic modality. The results show enrichment of genes previously selected for clinical development programs, as well as genome-wide significance in phenome-wide association studies.
COMMUNICATIONS BIOLOGY
(2022)
Article
Chemistry, Medicinal
Samuel Genheden, Per-Ola Norrby, Ola Engkvist
Summary: We present AiZynthTrain Python package that trains synthesis models in a robust, reproducible, and extensible way. It includes two pipelines for creating one-step retrosynthesis and RingBreaker models, which can be easily integrated into retrosynthesis software. By training on publicly available reaction data from USPTO, these are the first reproducible end-to-end retrosynthesis models. The pipeline demonstrates improved RingBreaker performance and robustness when trained on a more diverse proprietary dataset. This framework is expected to be expanded with other synthesis models in the future.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2023)
Article
Chemistry, Medicinal
Jonathan G. M. Conn, James W. Carter, Justin J. A. Conn, Vigneshwari Subramanian, Andrew Baxter, Ola Engkvist, Antonio Llinas, Ekaterina L. Ratkova, Stephen D. Pickett, James L. McDonagh, David S. Palmer
Summary: Accurate prediction of solubility is highly desirable in the field of chemical sciences. In 2019, the American Chemical Society organized a Solubility Challenge to evaluate the state of the art in this area. This article describes the development of two models submitted to the challenge, based on computationally inexpensive molecular descriptors and traditional machine learning algorithms. The performance of these models is compared to more advanced algorithms and larger training data sets, revealing potential areas for improvement in solubility prediction models.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2023)
Article
Chemistry, Physical
Morag Rose Hunter, Lili Cui, Benjamin Thomas Porebski, Sara Pereira, Silvia Sonzini, Uchechukwu Odunze, Preeti Iyer, Ola Engkvist, Rebecca Louise Lloyd, Samantha Peel, Alan Sabirsh, Douglas Ross-Thriepland, Arwyn Tomos Jones, Arpan Shailesh Desai
Summary: This study utilized siRNA and small molecule profiling approaches along with advanced imaging and machine learning techniques to explore the mechanism of lipid nanoparticle delivery of mRNA. By analyzing data-rich phenotypic fingerprints extracted from images, the researchers identified fluid-phase endocytosis as a productive cellular entry route for enhanced delivery. The re-engineered nanoparticle targeting macropinocytosis significantly improved mRNA delivery in vitro and in vivo.
Correction
Biology
Arwa Raies, Ewa Tulodziecka, James Stainer, Lawrence Middleton, Ryan S. S. Dhindsa, Pamela Hill, Ola Engkvist, Andrew R. R. Harper, Slave Petrovski, Dimitrios Vitsios
COMMUNICATIONS BIOLOGY
(2023)
Article
Chemistry, Multidisciplinary
Simon Viet Johansson, Morteza Haghir Chehreghani, Ola Engkvist, Alexander Schliep
Summary: Artificial intelligence (AI) offers new approaches to design compounds in drug discovery, such as suggesting new molecular structures or optimizing existing leads. However, the lack of high-quality data sets limits the effectiveness of AI methods. This study proposes a framework for designing combinatorial libraries using a molecular generative model and optimization algorithms. Simulation experiments show that near-optimal library designs can be achieved even without synthesizing additional building blocks.
Article
Chemistry, Multidisciplinary
Jeff Guo, Franziska Knuth, Christian Margreitter, Jon Paul Janet, Kostas Papadopoulos, Ola Engkvist, Atanas Patronov
Summary: This work presents Link-INVENT as an extension to the existing de novo molecular design platform REINVENT. Illustrative examples demonstrate its applications in fragment linking, scaffold hopping, and PROTAC design case studies. Link-INVENT, with the help of reinforcement learning, allows the agent to generate favorable linkers that connect molecular subunits satisfying diverse objectives, making it practical for real-world drug discovery projects.
Correction
Computer Science, Artificial Intelligence
Jeff Guo, Vendy Fialkova, Juan Diego Arango, Christian Margreitter, Jon Paul Janet, Kostas Papadopoulos, Ola Engkvist, Atanas Patronov
NATURE MACHINE INTELLIGENCE
(2022)
Article
Computer Science, Artificial Intelligence
Jeff Guo, Vendy Fialkova, Juan Diego Arango, Christian Margreitter, Jon Paul Janet, Kostas Papadopoulos, Ola Engkvist, Atanas Patronov
Summary: Guo and colleagues introduce curriculum learning extension to REINVENT, a de novo molecular design framework, which improves training efficiency by providing increasingly difficult problems over epochs.
NATURE MACHINE INTELLIGENCE
(2022)