Opportunities and Challenges for Machine Learning-Assisted Enzyme Engineering

ACS Central Science, Journal Year: 2024, Volume and Issue: 10(2), P. 226 - 241

Published: Feb. 5, 2024

Enzymes can be engineered at the level of their amino acid sequences to optimize key properties such as expression, stability, substrate range, and catalytic efficiency-or even unlock new activities not found in nature. Because search space possible proteins is vast, enzyme engineering usually involves discovering an starting point that has some desired activity followed by directed evolution improve its "fitness" for a application. Recently, machine learning (ML) emerged powerful tool complement this empirical process. ML models contribute (1) discovery functional annotation known protein or generating novel with functions (2) navigating fitness landscapes optimization mappings between associated values. In Outlook, we explain how complements discuss future potential improved outcomes.

Language: Английский

---

Machine Learning and Deep Learning in Synthetic Biology: Key Architectures, Applications, and Challenges

ACS Omega, Journal Year: 2024, Volume and Issue: 9(9), P. 9921 - 9945

Published: Feb. 19, 2024

Machine learning (ML), particularly deep (DL), has made rapid and substantial progress in synthetic biology recent years. Biotechnological applications of biosystems, including pathways, enzymes, whole cells, are being probed frequently with time. The intricacy interconnectedness biosystems make it challenging to design them the desired properties. ML DL have a synergy biology. Synthetic can be employed produce large data sets for training models (for instance, by utilizing DNA synthesis), ML/DL inform example, generating new parts or advising unrivaled experiments perform). This potential recently been brought light research at intersection engineering through achievements like novel biological components, best experimental design, automated analysis microscopy data, protein structure prediction, biomolecular implementations ANNs (Artificial Neural Networks). I divided this review into three sections. In first section, describe predictive basics along myriad biology, especially activity proteins, metabolic pathways. second fundamental architectures their Finally, different challenges causing hurdles solutions.

Language: Английский

---

Automated in vivo enzyme engineering accelerates biocatalyst optimization

Nature Communications, Journal Year: 2024, Volume and Issue: 15(1)

Published: April 24, 2024

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

Language: Английский

---

Perspectives on Computational Enzyme Modeling: From Mechanisms to Design and Drug Development

ACS Omega, Journal Year: 2024, Volume and Issue: unknown

Published: Feb. 8, 2024

Understanding enzyme mechanisms is essential for unraveling the complex molecular machinery of life. In this review, we survey field computational enzymology, highlighting key principles governing and discussing ongoing challenges promising advances. Over years, computer simulations have become indispensable in study mechanisms, with integration experimental exploration now established as a holistic approach to gain deep insights into enzymatic catalysis. Numerous studies demonstrated power characterizing reaction pathways, transition states, substrate selectivity, product distribution, dynamic conformational changes various enzymes. Nevertheless, significant remain investigating multistep reactions, large-scale changes, allosteric regulation. Beyond mechanistic studies, modeling has emerged an tool computer-aided design rational discovery covalent drugs targeted therapies. Overall, design/engineering drug development can greatly benefit from our understanding detailed enzymes, such protein dynamics, entropy contributions, allostery, revealed by studies. Such convergence different research approaches expected continue, creating synergies research. This outlining ever-expanding research, aims provide guidance future directions facilitate new developments important evolving field.

Language: Английский

---

Unlocking the potential of enzyme engineering via rational computational design strategies

Biotechnology Advances, Journal Year: 2024, Volume and Issue: 73, P. 108376 - 108376

Published: May 11, 2024

Language: Английский

---

Computational and Machine Learning Methods for CO₂ Capture Using Metal–Organic Frameworks

ACS Nano, Journal Year: 2024, Volume and Issue: 18(35), P. 23842 - 23875

Published: Aug. 22, 2024

Machine learning (ML) using data sets of atomic and molecular force fields (FFs) has made significant progress provided benefits in the chemistry material science. This work examines interactions between materials computational science at scales for metal-organic framework (MOF) adsorbent development toward carbon dioxide (CO

Language: Английский

---

Simplification of Corticosteroids Biosynthetic Pathway by Engineering P450BM3

ACS Catalysis, Journal Year: 2024, Volume and Issue: 14(6), P. 4117 - 4129

Published: March 4, 2024

Synthesis of corticosteroids, particularly hydrocortisone, is challenging owing to the complex network requiring pairing cytochrome P450s with P450 reductase (CPR) for achieving regionally selective hydroxylation modifications at multiple sites. Herein, we engineered a self-sufficient P450BM3 (CYP102A1 from Bacillus megaterium) effectively reducing traditionally complex, multienzyme cascade process (three steps and six enzymes) hydrocortisone synthesis progesterone (PG) simplified two-step involving least two enzymes. Driven by computational simulation-guided substrate access channel heme center pocket engineering, series variants were gradually designed ability catalyze C16β, C17α, C21, C17α/21 oxidation PG C11α cortexolone (c). Subsequently, molecular dynamics simulations an oxy-ferrous model revealed that glycine mutations residues are repulsive allow more stable exposure above Fe═O. Finally, developed employed construct efficient Escherichia coli catalytic systems, which further achieved 11α/β-hydrocortisone (f/e) production in one pot 1 g/L molar conversion rate 81 84% (912 955 mg/L), respectively. Thus, this study provides feasible strategies simplifying biosynthetic biocatalysts steroidal pharmaceutical production.

Language: Английский

---

Practical Machine Learning-Assisted Design Protocol for Protein Engineering: Transaminase Engineering for the Conversion of Bulky Substrates

ACS Catalysis, Journal Year: 2024, Volume and Issue: 14(9), P. 6462 - 6469

Published: April 12, 2024

Protein engineering is essential for improving the catalytic performance of enzymes applications in biocatalysis, which machine learning provides an emerging approach variant design. Transaminases are powerful biocatalysts stereoselective synthesis chiral amines but one major challenge their limited substrate scope. We present a general and practical design protocol protein to combine advantages three strategies, including directed evolution, rational design, learning, demonstrate application transaminases with higher activity toward bulky substrates. A high-quality data set was obtained by selected key positions, then applied create model transaminase activity. This data-assisted optimized variants, showed improved (up 3-fold over parent) substrates, maintaining enantioselectivity starting enzyme scaffold as well enantiomeric excess >99%ee).

Language: Английский

---

Best practices for machine learning in antibody discovery and development

Drug Discovery Today, Journal Year: 2024, Volume and Issue: 29(7), P. 104025 - 104025

Published: May 17, 2024

In the past 40 years, therapeutic antibody discovery and development have advanced considerably, with machine learning (ML) offering a promising way to speed up process by reducing costs number of experiments required. Recent progress in ML-guided design (D&D) has been hindered diversity data sets evaluation methods, which makes it difficult conduct comparisons assess utility. Establishing standards guidelines will be crucial for wider adoption ML advancement field. This perspective critically reviews current practices, highlights common pitfalls proposes method various ML-based techniques D&D. Addressing challenges across process, best practices are recommended each stage enhance reproducibility progress.

Language: Английский

---

MIFS: An adaptive multipath information fused self-supervised framework for drug discovery

Neural Networks, Journal Year: 2025, Volume and Issue: 184, P. 107088 - 107088

Published: Jan. 2, 2025

Language: Английский

---