Enzyme Engineering Database (EnzEngDB): a platform for sharing and interpreting sequence-function relationships across protein engineering campaigns.

In recent years, data visualization technology has received widespread attention, offering users a way to gain an in-depth understanding and effective exploration of various datasets by visualizing their composition and characteristics. Simultaneously, with the continuous development of hardware and algorithms, artificial intelligence (AI) has permeated every aspect of contemporary social life. Some studies have proposed the use of intelligent visualization tools that leverage AI to achieve significant results in data analysis, decision support, and automated reporting. However, there is a lack of systematic work to summarize and analyze these literature. In this paper, we conduct a comprehensive review of AI-based intelligent visualization tools and their applications in various domains. Firstly, we introduced the foundational concepts of AI and data visualization and explored the relationship between these two. we provided an overview of prevalent AI-based intelligent visualization tools along with their diverse applications. Our discussion encompassed renowned tools like Tableau, Power BI, QlikView, D3.js, IBM Watson Analytics, and FusionCharts, delving into their distinctive features, functionalities, and practical implementations. Furthermore, we have explored the advantages and challenges of AI-integrated visualization. The advantages include enhanced data understanding and insights, as well as automation of analysis and visualization generation. Some challenges coexist in the meantime, including data quality and accuracy, privacy and security concerns, and user experience and usability issues. Based on our findings, we have added a great deal to the corpus of knowledge by giving a thorough rundown of visualization tools with AI integration and their possible uses. We have also identified existing issues and limitations that need to be addressed for further advancements in this field. Lastly, we have proposed future research directions and development trends to guide researchers and practitioners in fully harnessing the potential of AI and visualization.

Protein engineering, the process by which novel proteins with desired properties are developed, has grown by leaps and bounds since the first examples of protein mutagenesis were described over three decades ago. In the last 30 years, protein engineers have successfully created a wide range of proteins tailored to specific industrial, medical, and research applications. Yet challenges remain that prevent the engineering of complex protein functions on demand. With the aim of creating a forum for the protein engineering community, the inaugural Protein Engineering Canada Conference was held on June 20–22, 2014, in Ottawa, Canada. Over the course of the 2-day conference, 115 protein scientists from over 30 institutions representing 5 countries shared ideas, networked with colleagues, and learned about exciting new research covering a broad array of topics such as enzyme engineering, computational protein design, X-ray crystallography, protein NMR, molecular modeling, and protein evolution. This special issue highlights the breadth of research that was covered during the Protein Engineering Canada Conference by including previously unpublished research articles submitted by speakers as well as some selections that were submitted directly to Protein Science.

Esters are vital compounds with wide-ranging applications in food, fragrance, and pharmaceuticals, yet their sustainable supply increasingly relies on microbial biosynthesis. Acyltransferases (ATFs), particularly BAHD family members, catalyze the esterification of alcohols with acyl-CoA donors, a central step in generating structurally diverse esters in vivo. This review highlights recent progress in ATF-mediated microbial biosynthesis of three representative classes: short-chain fatty acid esters (SCFAEs), monoterpenyl esters (MTEs), and aromatic esters (AEs). Advances in enzyme discovery, protein engineering, and metabolic pathway reprogramming have expanded substrate scope, improved catalytic efficiency, and enabled titers from milligram to gram scale. Nonetheless, challenges remain in achieving high activity, selectivity, and stereochemical fidelity under industrial conditions. We summarize strategies including co-culture systems, non-natural precursor biosynthetic pathways, high-throughput screening, and rational enzyme engineering. Future perspectives emphasize developing a precision precursor supply for ATF-mediated esterification, elucidating ATF structure-function relationships, and optimizing downstream processing to achieve efficient, scalable, and sustainable microbial production of diverse natural esters.

Biotechnology-enabled plastic valorization across enzymatic, microbial, and hybrid conversion systems: A review.

Advances and opportunities for the design of self-sufficient and spatially organized cell-free biocatalytic systems

In this dispensation of the fourth industrial revolution, protein engineering has become a popular approach for increasing enzymatic activity, stability, and titer in the biosynthesis of natural products. This is attributed to its numerous advantages (over direct isolation from plants or via chemical synthesis), including decreasing or eliminating reaction byproducts, high precision, moderate handling of intricate and chemically unstable chemicals, overall reusability, and cost efficiency. Recently, protein engineering tools have advanced to redesign and enhance natural product biosynthesis. These methods include direct evolution, substrate engineering, medium engineering, enzyme engineering and immobilization, structure-assisted protein engineering, and advanced computational. Recent successes in implementing these emerging protein engineering technologies were critically discussed in this article. Also, the advantages, limitations, and applications in industrial and medical biotechnology were discussed. Last, future research directions and potential were also highlighted.

Phytases are important industrial enzymes able to catalyze the release of up to six phosphates from phytate in a stepwise hydrolysis reaction. Phytases are almost exclusively used as a feed supplement. However, phytases are also used in human nutrition, food processing, non-food industrial products, and emerging applications like enzymatic phosphate recovery from renewable resources. Phytate, the main phosphorus storage form in seeds, and its hydrolysis products act as a chelator and reduce protein and mineral bioavailability in intestinal absorption. Full phosphate hydrolysis from the common storage compound phytate remains a challenge. Phytate hydrolysis patterns of tailored phytases and their protein engineering campaigns are discussed. The aim of our review is to give an overview on developed and emerging application areas (animal nutrition, food processing, and environmental resource management) and thereby generate an awareness for the importance of phosphorus stewardship in a circular bioeconomy. Emphasis will be given to processes using organic-bound phosphorus and related recycling strategy of this valuable resource. In detail, the main challenge in designing phytases to completely hydrolyze phosphate from phytate to inositol and the need for engineering campaigns to broaden their industrial use are described.

Several feature extraction approaches for QSPR modelling in Cheminformatics are discussed in this paper. In particular, this work is focused on the use of these strategies for predicting mechanical properties, which are relevant for the design of polymeric materials. The methodology analysed in this study employs a feature learning method that uses a quantification process of 2D structural characterization of materials with the autoencoder method. Alternative QSPR models inferred for tensile strength at break (a well-known mechanical property of polymers) are presented. These alternative models are contrasted to QSPR models obtained by feature selection technique by using accuracy measures and a visual analytic tool. The results show evidence about the benefits of combining feature learning approaches with feature selection methods for the design of QSPR models.

Introduction Advances in large language models (LLMs) have opened new avenues for healthcare applications. Recently, ChatGPT-4 successfully achieved the pass mark >80% in 5 of 10 sleep medicine examination domains, indicating a strong foundational knowledge of sleep medicine. However, the ability to answer USMLE-type multiple-choice questions may not equate to the capacity to offer accurate and comprehensive answers to clinical queries or real-world case scenarios. Current literature exhibits a significant gap in validation research examining the clinical utility of LLMs in evidence-based medicine practice. This investigation aims to evaluate the potential usefulness and reliability of LLMs as adjunctive clinical decision-support tools (namely, “curbside consultants”) in sleep medicine. Methods Six clinical sleep queries and six case scenarios were presented to 6 LLMs including 3 general-purpose LLMs (ChatGPT-4o, Gemini-1.5-Pro, and Llama-3.1-405B) and 3 medical-specialized LLMs (OpenEvidence, Clara AI, and MediGPT). Performance assessment was conducted independently utilizing two independently developed 5-point Likert scales evaluating two primary domains: answer content (accuracy, relevance, comprehensiveness/depth, clarity/coherence, and unique insightfulness) and reference quality (accuracy, relevance, currentness, comprehensiveness/depth, and searchability). Benchmark answers were established through consensus among four sleep medicine specialists. Analysis of variance was used to compare the performance of the LLMs. Results The medical LLMs demonstrated superior overall performance compared to the general LLMs (P< 0.001). The primary distinction was observed in reference quality metrics, where medical LLMs significantly outperformed general LLMs across all parameters: accuracy, relevance, and searchability (p< 0.001), currentness (p=0.001), and comprehensiveness/depth (p=0.010). Notably, Open Evidence achieved the highest reference quality (p< 0.001). In contrast, the answer content analysis revealed no significant overall differences between medical and general LLMs (p=0.659). Most answer content-related metrics, including accuracy, relevance, or unique insightfulness, did not differ significantly (p>0.050). An exception was clarify/coherence, where medical LLMs were superior to general LLMs (p=0.030). Furthermore, MediGPT and ChatGPT-4o displayed better content comprehensiveness/depth relative to other LLMs (p< 0.001). These two LLMs exhibited comparable performance in overall metrics (p=0.615), answer contents (p=0.922), and reference quality (p=0.621). Conclusion Both medical-specialized and general-purpose LLMs show promise as adjunctive decision-support tools in clinical practice. However, substantial improvements in reference quality are critically needed across most LLM platforms. Support (if any)

Age-related macular degeneration (AMD) represents a significant clinical concern, particularly in aging populations, and recent advancements in artificial intelligence (AI) have catalyzed substantial research interest in this domain. Despite the growing body of literature, there remains a need for a comprehensive, quantitative analysis to delineate key trends and emerging areas in the field of AI applications in AMD. This bibliometric analysis sought to systematically evaluate the landscape of AI-focused research on AMD to illuminate publication patterns, influential contributors, and focal research trends. Using the Web of Science Core Collection (WoSCC), a search was conducted to retrieve relevant publications from 1992 to 2023. This analysis involved an array of bibliometric indicators to map the evolution of AI research in AMD, assessing parameters such as publication volume, national/regional and institutional contributions, journal impact, author influence, and emerging research hotspots. Visualization tools, including Bibliometrix, CiteSpace and VOSviewer, were employed to generate comprehensive assessments of the data. A total of 1,721 publications were identified, with the USA leading in publication output and the University of Melbourne as the most prolific institution. The journal Investigative Ophthalmology & Visual Science published the highest number of articles, and Schmidt-Eerfurth emerged as the most active author. Keyword and clustering analyses, along with citation burst detection, revealed three distinct research stages within the field from 1992 to 2023. Presently, research efforts are concentrated on developing deep learning (DL) models for AMD diagnosis and progression prediction. Prominent emerging themes include early detection, risk stratification, and treatment efficacy prediction. The integration of large language models (LLMs) and vision-language models (VLMs) for enhanced image processing also represents a novel research frontier. This bibliometric analysis provides a structured overview of prevailing research trends and emerging directions in AI applications for AMD. These findings furnish valuable insights to guide future research and foster collaborative advancements in this evolving field.

Evaluating large language models (LLMs) presents unique challenges. While automatic side-by-side evaluation, also known as LLM-as-a-judge, has become a promising solution, model developers and researchers face difficulties with scalability and interpretability when analyzing these evaluation outcomes. To address these challenges, we introduce LLM Comparator, a new visual analytics tool designed for side-by-side evaluations of LLMs. This tool provides analytical workflows that help users understand when and why one LLM outperforms or underperforms another, and how their responses differ. Through close collaboration with practitioners developing LLMs at Google, we have iteratively designed, developed, and refined the tool. Qualitative feedback from these users highlights that the tool facilitates in-depth analysis of individual examples while enabling users to visually overview and flexibly slice data. This empowers users to identify undesirable patterns, formulate hypotheses about model behavior, and gain insights for model improvement. LLM Comparator has been integrated into Google's LLM evaluation platforms and open-sourced.

Transcription factor (TF)-based biosensors are very desirable reagents for high-throughput enzyme and strain engineering campaigns. Despite their potential, they are often difficult to deploy effectively as the small molecules being detected can leak out of high-producer cells, into low-producer cells, and activate the biosensor therein. This crosstalk leads to the overrepresentation of false-positive/cheater cells in the enriched population. While the host cell can be engineered to minimize crosstalk (e.g., by deleting responsible transporters), this is not easily applicable to all molecules of interest, particularly those that can diffuse passively. One such biosensor recently reported for trans-cinnamic acid (tCA) suffers from crosstalk when used for phenylalanine ammonia-lyase (PAL) enzyme engineering by directed evolution. We report that desensitizing the biosensor (i.e., increasing the limit of detection) suppresses cheater population enrichment. Furthermore, we show that, if we couple the biosensor-based screen with an orthogonal prescreen that eliminates a large fraction of true negatives, we can successfully reduce the cheater population during the fluorescence-activated cell sorting. Using the approach developed here, we were successfully able to isolate PAL variants with ∼70% higher kcat after a single sort. These mutants have tremendous potential in phenylketonuria (PKU) treatment and flavonoid production.

Large language models for electronic health records in pediatric and surgical care: A systematic review.

Protein engineering of representative hydrolytic enzymes for remediation of organophosphates

The synthesis of secondary and tertiary amines through the reductive amination of carbonyl compounds is one of the most significant reactions in synthetic chemistry. Asymmetric reductive amination for the formation of chiral amines, which are required for the synthesis of pharmaceuticals and other bioactive molecules, is often achieved through transition metal catalysis, but biocatalytic methods of chiral amine production have also been a focus of interest owing to their selectivity and sustainability. The discovery of asymmetric reductive amination by imine reductase (IRED) and reductive aminase (RedAm) enzymes has served as the starting point for a new industrial approach to the production of chiral amines, leading from laboratory-scale milligram transformations to ton-scale reactions that are now described in the public domain. In this perspective we trace the development of the IRED-catalyzed reductive amination reaction from its discovery to its industrial application on kg to ton scale. In addition to surveying examples of the synthetic chemistry that has been achieved with the enzymes, the contribution of structure and protein engineering to the understanding of IRED-catalyzed reductive amination is described, and the consequent benefits for activity, selectivity and stability in the design of process suitable catalysts.