Deep learning driven biosynthetic pathways navigation for natural products with BioNavi-NP

Planning biosynthetic pathways of target molecules based on metabolic reaction prediction and AND-OR tree search

Bioretrosynthesis problem is to predict synthetic routes using substrates for given natural products (NPs). However, the huge number of metabolic reactions leads to a combinatorial explosion of searching space, which is high time-consuming and costly. Here, we propose a framework called BioRetro to predict bioretrosynthesis pathways using a one-step bioretrosynthesis network, termed HybridMLP combined with AND-OR tree heuristic search. The HybridMLP predicts precursors that will produce the target NPs, while the AND-OR tree generates the iterative multi-step biosynthetic pathways. The one-step bioretrosynthesis prediction experiments are conducted on MetaNetX dataset by using HybridMLP, which achieves 46.5%, 74.6%, 81.6% in terms of the top-1, top-5, top-10 accuracies. The great performance demonstrates the effectiveness of HybridMLP in one-step bioretrosynthesis. Besides, the evaluation of two benchmark datasets reveals that BioRetro can significantly improve the speed and success rate in predicting biosynthesis pathways. In addition, the BioRetro is further shown to find the synthetic pathway of compounds, such as ginsenoside F1 with the same substrates as reported but different enzymes, which may be the novel potential enzyme to have better catalytic performance.

Advancing microbial production through artificial intelligence-aided biology

Microbial cell factories (MCFs) have been leveraged to construct sustainable platforms for value-added compound production. To optimize metabolism and reach optimal productivity, synthetic biology has developed various genetic devices to engineer microbial systems by gene editing, high-throughput protein engineering, and dynamic regulation. However, current synthetic biology methodologies still rely heavily on manual design, laborious testing, and exhaustive analysis. The emerging interdisciplinary field of artificial intelligence (AI) and biology has become pivotal in addressing the remaining challenges. AI-aided microbial production harnesses the power of processing, learning, and predicting vast amounts of biological data within seconds, providing outputs with high probability. With well-trained AI models, the conventional Design-Build-Test (DBT) cycle has been transformed into a multidimensional Design-Build-Test-Learn-Predict (DBTLP) workflow, leading to significantly improved operational efficiency and reduced labor consumption. Here, we comprehensively review the main components and recent advances in AI-aided microbial production, focusing on genome annotation, AI-aided protein engineering, artificial functional protein design, and AI-enabled pathway prediction. Finally, we discuss the challenges of integrating novel AI techniques into biology and propose the potential of large language models (LLMs) in advancing microbial production.

Hydrophobic substrate binding pocket remodeling of echinocandin B deacylase based on multi-dimensional rational design

Actinoplanes utahensis deacylase (AAC)-catalyzed deacylation of echinocandin B (ECB) is a promising method for the synthesis of anidulafungin, the newest of the echinocandin antifungal agents. However, the low activity of AAC significantly limits its practical application. In this work, we have devised a multi-dimensional rational design strategy for AAC, conducting separate analyses on the substrate-binding pocket's volume, curvature, and length. Furthermore, we quantitatively analyzed substrate properties, particularly on hydrophilic and hydrophobic. Accordingly, we tailored the linoleic acid-binding pocket of AAC to accommodate the extended long lipid chain of ECB. By fine-tuning the key residues, the resulting AAC mutants can accommodate the ECB lipid chain with a lower curvature binding pocket. The D53A/I55F/G57M/F154L/Q661L mutant (MT) displayed 331 % higher catalytic efficiency than the wild-type (WT) enzyme. The MT product conversion was 94.6 %, reaching the highest reported level. Utilizing a multi-dimensional rational design for a customized mutation strategy of the substrate-binding pocket is an effective approach to enhance the catalytic efficiency of enzymes in handling complicated substrates.

Automated in vivo enzyme engineering accelerates biocatalyst optimization

Achieving cost-competitive bio-based processes requires development of stable and selective biocatalysts. Their realization through in vitro enzyme characterization and engineering is mostly low throughput and labor-intensive. Therefore, strategies for increasing throughput while diminishing manual labor are gaining momentum, such as in vivo screening and evolution campaigns. Computational tools like machine learning further support enzyme engineering efforts by widening the explorable design space. Here, we propose an integrated solution to enzyme engineering challenges whereby ML-guided, automated workflows (including library generation, implementation of hypermutation systems, adapted laboratory evolution, and in vivo growth-coupled selection) could be realized to accelerate pipelines towards superior biocatalysts.

A general model for predicting enzyme functions based on enzymatic reactions

Accurate prediction of the enzyme comission (EC) numbers for chemical reactions is essential for the understanding and manipulation of enzyme functions, biocatalytic processes and biosynthetic planning. A number of machine leanring (ML)-based models have been developed to classify enzymatic reactions, showing great advantages over costly and long-winded experimental verifications. However, the prediction accuracy for most available models trained on the records of chemical reactions without specifying the enzymatic catalysts is rather limited. In this study, we introduced BEC-Pred, a BERT-based multiclassification model, for predicting EC numbers associated with reactions. Leveraging transfer learning, our approach achieves precise forecasting across a wide variety of Enzyme Commission (EC) numbers solely through analysis of the SMILES sequences of substrates and products. BEC-Pred model outperformed other sequence and graph-based ML methods, attaining a higher accuracy of 91.6%, surpassing them by 5.5%, and exhibiting superior F1 scores with improvements of 6.6% and 6.0%, respectively. The enhanced performance highlights the potential of BEC-Pred to serve as a reliable foundational tool to accelerate the cutting-edge research in synthetic biology and drug metabolism. Moreover, we discussed a few examples on how BEC-Pred could accurately predict the enzymatic classification for the Novozym 435-induced hydrolysis and lipase efficient catalytic synthesis. We anticipate that BEC-Pred will have a positive impact on the progression of enzymatic research.

Efficient retrosynthetic planning with MCTS exploration enhanced A* search

Retrosynthetic planning, which aims to identify synthetic pathways for target molecules from starting materials, is a fundamental problem in synthetic chemistry. Computer-aided retrosynthesis has made significant progress, in which heuristic search algorithms, including Monte Carlo Tree Search (MCTS) and A* search, have played a crucial role. However, unreliable guiding heuristics often cause search failure due to insufficient exploration. Conversely, excessive exploration also prevents the search from reaching the optimal solution. In this paper, MCTS exploration enhanced A* (MEEA*) search is proposed to incorporate the exploratory behavior of MCTS into A* by providing a look-ahead search. Path consistency is adopted as a regularization to improve the generalization performance of heuristics. Extensive experimental results on 10 molecule datasets demonstrate the effectiveness of MEEA*. Especially, on the widely used United States Patent and Trademark Office (USPTO) benchmark, MEEA* achieves a 100.0% success rate. Moreover, for natural products, MEEA* successfully identifies bio-retrosynthetic pathways for 97.68% test compounds.

High-throughput prediction of enzyme promiscuity based on substrate-product pairs

The screening of enzymes for catalyzing specific substrate-product pairs is often constrained in the realms of metabolic engineering and synthetic biology. Existing tools based on substrate and reaction similarity predominantly rely on prior knowledge, demonstrating limited extrapolative capabilities and an inability to incorporate custom candidate-enzyme libraries. Addressing these limitations, we have developed the Substrate-product Pair-based Enzyme Promiscuity Prediction (SPEPP) model. This innovative approach utilizes transfer learning and transformer architecture to predict enzyme promiscuity, thereby elucidating the intricate interplay between enzymes and substrate-product pairs. SPEPP exhibited robust predictive ability, eliminating the need for prior knowledge of reactions and allowing users to define their own candidate-enzyme libraries. It can be seamlessly integrated into various applications, including metabolic engineering, de novo pathway design, and hazardous material degradation. To better assist metabolic engineers in designing and refining biochemical pathways, particularly those without programming skills, we also designed EnzyPick, an easy-to-use web server for enzyme screening based on SPEPP. EnzyPick is accessible at http://www.biosynther.com/enzypick/.

Enabling pathway design by multiplex experimentation and machine learning

The remarkable metabolic diversity observed in nature has provided a foundation for sustainable production of a wide array of valuable molecules. However, transferring the biosynthetic pathway to the desired host often runs into inherent failures that arise from intermediate accumulation and reduced flux resulting from competing pathways within the host cell. Moreover, the conventional trial and error methods utilized in pathway optimization struggle to fully grasp the intricacies of installed pathways, leading to time-consuming and labor-intensive experiments, ultimately resulting in suboptimal yields. Considering these obstacles, there is a pressing need to explore the enzyme expression landscape and identify the optimal pathway configuration for enhanced production of molecules. This review delves into recent advancements in pathway engineering, with a focus on multiplex experimentation and machine learning techniques. These approaches play a pivotal role in overcoming the limitations of traditional methods, enabling exploration of a broader design space and increasing the likelihood of discovering optimal pathway configurations for enhanced production of molecules. We discuss several tools and strategies for pathway design, construction, and optimization for sustainable and cost-effective microbial production of molecules ranging from bulk to fine chemicals. We also highlight major successes in academia and industry through compelling case studies.

Metabolomics of natural samples: A tutorial review on the latest technologies

Metabolomics is the study of metabolites present in a living system. It is a rapidly growing field aimed at discovering novel compounds, studying biological processes, diagnosing diseases, and ensuring the quality of food products. Recently, the analysis of natural samples has become important to explore novel bioactive compounds and to study how environment and genetics affect living systems. Various metabolomics techniques, databases, and data analysis tools are available for natural sample metabolomics. However, choosing the right method can be a daunting exercise because natural samples are heterogeneous and require untargeted approaches. This tutorial review aims to compile the latest technologies to guide an early-career scientist on natural sample metabolomics. First, different extraction methods and their pros and cons are reviewed. Second, currently available metabolomics databases and data analysis tools are summarized. Next, recent research on metabolomics of milk, honey, and microbial samples is reviewed. Finally, after reviewing the latest trends in technologies, a checklist is presented to guide an early-career researcher on how to design a metabolomics project. In conclusion, this review is a comprehensive resource for a researcher planning to conduct their first metabolomics analysis. It is also useful for experienced researchers to update themselves on the latest trends in metabolomics.

Coupling chemistry and biology for the synthesis of advanced bioproducts

Chemical and biological syntheses can both lead to a myriad of compounds. Biology enables us to harness the metabolism of microbial cell factories to produce key target molecules from renewable biomass-derived substrates. Although bio-based feedstocks are sustainably sourced and more benign than the rapidly depleting fossil fuels that chemical processes have historically relied on, limiting pathways solely to biological reactions may not equate to a greener process overall. In fact, bioreactors rely on substantial quantities of water and can be inefficient since organisms typically operate around ambient conditions and are sensitive to perturbations in their environment. Hybridizing biosynthetic pathways with green chemistry can instead be a more potent strategy to reduce our net manufacturing footprint. Emerging chemistries have demonstrated considerable success in performing complex transformations on biological feedstocks without significant solvent use. Many of these transformations would be too slow to perform enzymatically or infeasible altogether. Here, we put forth the concept that by carefully considering the merits and drawbacks of synthetic biology and chemistry as well as one's own use case, there exist many opportunities for coupling the two. Merging these syntheses can unlock a wider suite of functional group transformations, thereby enabling future manufacturing processes to sustainably access a larger space of valuable, platform chemicals.

Machine Learning-Guided Protein Engineering

Recent progress in engineering highly promising biocatalysts has increasingly involved machine learning methods. These methods leverage existing experimental and simulation data to aid in the discovery and annotation of promising enzymes, as well as in suggesting beneficial mutations for improving known targets. The field of machine learning for protein engineering is gathering steam, driven by recent success stories and notable progress in other areas. It already encompasses ambitious tasks such as understanding and predicting protein structure and function, catalytic efficiency, enantioselectivity, protein dynamics, stability, solubility, aggregation, and more. Nonetheless, the field is still evolving, with many challenges to overcome and questions to address. In this Perspective, we provide an overview of ongoing trends in this domain, highlight recent case studies, and examine the current limitations of machine learning-based methods. We emphasize the crucial importance of thorough experimental validation of emerging models before their use for rational protein design. We present our opinions on the fundamental problems and outline the potential directions for future research.

Deep learning for metabolic pathway design

The establishment of a bio-based circular economy is imperative in tackling the climate crisis and advancing sustainable development. In this realm, the creation of microbial cell factories is central to generating a variety of chemicals and materials. The design of metabolic pathways is crucial in shaping these microbial cell factories, especially when it comes to producing chemicals with yet-to-be-discovered biosynthetic routes. To aid in navigating the complexities of chemical and metabolic domains, computer-supported tools for metabolic pathway design have emerged. In this paper, we evaluate how digital strategies can be employed for pathway prediction and enzyme discovery. Additionally, we touch upon the recent strides made in using deep learning techniques for metabolic pathway prediction. These computational tools and strategies streamline the design of metabolic pathways, facilitating the development of microbial cell factories. Leveraging the capabilities of deep learning in metabolic pathway design is profoundly promising, potentially hastening the advent of a bio-based circular economy.

Artificial intelligence for natural product drug discovery

Developments in computational omics technologies have provided new means to access the hidden diversity of natural products, unearthing new potential for drug discovery. In parallel, artificial intelligence approaches such as machine learning have led to exciting developments in the computational drug design field, facilitating biological activity prediction and de novo drug design for molecular targets of interest. Here, we describe current and future synergies between these developments to effectively identify drug candidates from the plethora of molecules produced by nature. We also discuss how to address key challenges in realizing the potential of these synergies, such as the need for high-quality datasets to train deep learning algorithms and appropriate strategies for algorithm validation.

Multistep retrosynthesis combining a disconnection aware triple transformer loop with a route penalty score guided tree search

Computer-aided synthesis planning (CASP) aims to automatically learn organic reactivity from literature and perform retrosynthesis of unseen molecules. CASP systems must learn reactions sufficiently precisely to propose realistic disconnections, while avoiding overfitting to leave room for diverse options, and explore possible routes such as to allow short synthetic sequences to emerge. Herein we report an open-source CASP tool proposing original solutions to both challenges. First, we use a triple transformer loop (TTL) predicting starting materials (T1), reagents (T2), and products (T3) to explore various disconnection sites defined by combining systematic, template-based, and transformer-based tagging procedures. Second, we integrate TTL into a multistep tree search algorithm (TTLA) prioritizing sequences using a route penalty score (RPScore) considering the number of steps, their confidence score, and the simplicity of all intermediates along the route. Our approach favours short synthetic routes to commercial starting materials, as exemplified by retrosynthetic analyses of recently approved drugs.

Hosts engineering and in vitro enzymatic synthesis for the discovery of novel natural products and their derivatives

Novel natural products (NPs) and their derivatives are important sources for drug discovery, which have been broadly applied in the fields of agriculture, livestock, and medicine, making the synthesis of NPs and their derivatives necessarily important. In recent years, biosynthesis technology has received increasing attention due to its high efficiency in the synthesis of high value-added novel products and its advantages of green, environmental protection, and controllability. In this review, the technological advances of biosynthesis strategies in the discovery of novel NPs and their derivatives are outlined, with an emphasis on two areas of host engineering and in vitro enzymatic synthesis. In terms of hosts engineering, multiple microorganisms, including Streptomyces, Aspergillus, and Penicillium, have been used as the biosynthetic gene clusters (BGCs) provider and host strain for the expression of BGCs to discover new compounds over the past years. In addition, the use of in vitro enzymatic synthesis strategy to generate novel compounds such as triterpenoid saponins and flavonoids is also hereby described.

Developing TeroENZ and TeroMAP modules for the terpenome research platform TeroKit

Terpenoids and their derivatives are collectively known as the terpenome and are the largest class of natural products, whose biosynthesis refers to various kinds of enzymes. To date, there is no terpenome-related enzyme database, which is a desire for enzyme mining, metabolic engineering and discovery of new natural products related to terpenoids. In this work, we have constructed a comprehensive database called TeroENZ (http://terokit.qmclab.com/browse_enz.html) containing 13 462 enzymes involved in the terpenoid biosynthetic pathway, covering 2541 species and 4293 reactions reported in the literature and public databases. At the same time, we classify enzymes according to their catalytic reactions into cyclase, oxidoreductase, transferase, and so on, and also make a classification according to species. This meticulous classification is beneficial for users as it can be retrieved and downloaded conveniently. We also provide a computational module for isozyme prediction. Moreover, a module named TeroMAP (http://terokit.qmclab.com/browse_rxn.html) is also constructed to organize all available terpenoid enzymatic reactions into an interactive network by interfacing with the previously established database of terpenoid compounds, TeroMOL. Finally, all these databases and modules are integrated into the web server TeroKit (http://terokit.qmclab.com/) to shed light on the field of terpenoid research. Database URL http://terokit.qmclab.com/.

Metabolic engineering for sustainability and health

Bio-based production of chemicals and materials has attracted much attention due to the urgent need to establish sustainability and enhance human health. Metabolic engineering (ME) allows purposeful modification of cellular metabolic, regulatory, and signaling networks to achieve enhanced production of desired chemicals and degradation of environmentally harmful chemicals. ME has significantly progressed over the past 30 years through further integration of the strategies of synthetic biology, systems biology, evolutionary engineering, and data science aided by artificial intelligence. Here we review the field of ME from its emergence to the current state-of-the-art, highlighting its contribution to sustainable production of chemicals, health, and the environment through representative examples. Future challenges of ME and perspectives are also discussed.

Redesigning plant specialized metabolism with supervised machine learning using publicly available reactome data

The immense structural diversity of products and intermediates of plant specialized metabolism (specialized metabolites) makes them rich sources of therapeutic medicine, nutrients, and other useful materials. With the rapid accumulation of reactome data that can be accessible on biological and chemical databases, along with recent advances in machine learning, this review sets out to outline how supervised machine learning can be used to design new compounds and pathways by exploiting the wealth of said data. We will first examine the various sources from which reactome data can be obtained, followed by explaining the different machine learning encoding methods for reactome data. We then discuss current supervised machine learning developments that can be employed in various aspects to help redesign plant specialized metabolism.

Merging enzymatic and synthetic chemistry with computational synthesis planning

Synthesis planning programs trained on chemical reaction data can design efficient routes to new molecules of interest, but are limited in their ability to leverage rare chemical transformations. This challenge is acute for enzymatic reactions, which are valuable due to their selectivity and sustainability but are few in number. We report a retrosynthetic search algorithm using two neural network models for retrosynthesis-one covering 7984 enzymatic transformations and one 163,723 synthetic transformations-that balances the exploration of enzymatic and synthetic reactions to identify hybrid synthesis plans. This approach extends the space of retrosynthetic moves by thousands of uniquely enzymatic one-step transformations, discovers routes to molecules for which synthetic or enzymatic searches find none, and designs shorter routes for others. Application to (-)-Δ⁹ tetrahydrocannabinol (THC) (dronabinol) and R,R-formoterol (arformoterol) illustrates how our strategy facilitates the replacement of metal catalysis, high step counts, or costly enantiomeric resolution with more elegant hybrid proposals.

Achievements and perspectives of synthetic biology in botanical insecticides

Botanical insecticides are the origin of all insecticidal compounds. They have been widely used to control pests in crops for a long time. Currently, the commercial production of botanical insecticides extracted from plants is limited because of insufficient raw material supply. Synthetic biology is a promising and effective approach for addressing the current problems of the production of botanical insecticides. It is an emerging biological research hotspot in the field of botanical insecticides. However, the biosynthetic pathways of many botanical insecticides are not completely elucidated. On the other hand, the cytotoxicity of botanical pesticides and low efficiency of these biosynthetic enzymes in new hosts make it still challenging for their heterologous production. In the present review, we summarized the recent developments in the heterologous production of botanical insecticides, analyzed the current challenges, and discussed the feasible production strategies, focusing on elucidating biosynthetic pathways, enzyme engineering, host engineering, and cytotoxicity engineering. Looking to the future, synthetic biology promises to further advance heterologous production of more botanical pesticides.

Engineered Saccharomyces cerevisiae for the De Novo Biosynthesis of (-)-Menthol

Menthol, a high-value commodity monoterpenoid chemical, holds an important market share commercially because of its distinct functions. The menthol on the market mainly originates from plant extraction, which is facing challenges such as the seasonal fluctuations and long growth cycle of plants. Therefore, this study attempted to realize the de novo synthesis of menthol through microbial fermentation. First, through heterologous expression and subcellular localization observation, a synthetic route from glucose to (-)-menthol was successfully designed and constructed in Saccharomyces cerevisiae. Then, the mevalonate (MVA) pathway was enhanced, and the expression of farnesyl diphosphate synthase (ERG20) was dynamically regulated to improve the synthesis of D-limonene, a key precursor of (-)-menthol. Shake flask fermentation results showed that the D-limonene titer of the recombinant strain reached 459.59 mg/L. Next, the synthesis pathway from D-limonene to (-)-menthol was strengthened, and the fermentation medium was optimized. The (-)-menthol titer of 6.28 mg/L was obtained, implying that the de novo synthesis of menthol was successfully realized for the first time. This study provides a good foundation for the synthesis of menthol through microbial fermentation.