Rosetta and the Journey to Predict Proteins’ Structures, 20 Years on

AI-Driven Deep Learning Techniques in Protein Structure Prediction

Protein structure prediction is important for understanding their function and behavior. This review study presents a comprehensive review of the computational models used in predicting protein structure. It covers the progression from established protein modeling to state-of-the-art artificial intelligence (AI) frameworks. The paper will start with a brief introduction to protein structures, protein modeling, and AI. The section on established protein modeling will discuss homology modeling, ab initio modeling, and threading. The next section is deep learning-based models. It introduces some state-of-the-art AI models, such as AlphaFold (AlphaFold, AlphaFold2, AlphaFold3), RoseTTAFold, ProteinBERT, etc. This section also discusses how AI techniques have been integrated into established frameworks like Swiss-Model, Rosetta, and I-TASSER. The model performance is compared using the rankings of CASP14 (Critical Assessment of Structure Prediction) and CASP15. CASP16 is ongoing, and its results are not included in this review. Continuous Automated Model EvaluatiOn (CAMEO) complements the biennial CASP experiment. Template modeling score (TM-score), global distance test total score (GDT_TS), and Local Distance Difference Test (lDDT) score are discussed too. This paper then acknowledges the ongoing difficulties in predicting protein structure and emphasizes the necessity of additional searches like dynamic protein behavior, conformational changes, and protein-protein interactions. In the application section, this paper introduces some applications in various fields like drug design, industry, education, and novel protein development. In summary, this paper provides a comprehensive overview of the latest advancements in established protein modeling and deep learning-based models for protein structure predictions. It emphasizes the significant advancements achieved by AI and identifies potential areas for further investigation.

Enzyme-assisted high throughput sequencing of an expanded genetic alphabet at single base resolution

With just four building blocks, low sequence information density, few functional groups, poor control over folding, and difficulties in forming compact folds, natural DNA and RNA have been disappointing platforms from which to evolve receptors, ligands, and catalysts. Accordingly, synthetic biology has created "artificially expanded genetic information systems" (AEGIS) to add nucleotides, functionality, and information density. With the expected improvements seen in AegisBodies and AegisZymes, the task for synthetic biologists shifts to developing for expanded DNA the same analytical tools available to natural DNA. Here we report one of these, an enzyme-assisted sequencing of expanded genetic alphabet (ESEGA) method to sequence six-letter AEGIS DNA. We show how ESEGA analyses this DNA at single base resolution, and applies it to optimized conditions for six-nucleotide PCR, assessing the fidelity of various DNA polymerases, and extending this to AEGIS components with functional groups. This supports the renewed exploitation of expanded DNA alphabets in biotechnology.

HNetGO: protein function prediction via heterogeneous network transformer

Protein function annotation is one of the most important research topics for revealing the essence of life at molecular level in the post-genome era. Current research shows that integrating multisource data can effectively improve the performance of protein function prediction models. However, the heavy reliance on complex feature engineering and model integration methods limits the development of existing methods. Besides, models based on deep learning only use labeled data in a certain dataset to extract sequence features, thus ignoring a large amount of existing unlabeled sequence data. Here, we propose an end-to-end protein function annotation model named HNetGO, which innovatively uses heterogeneous network to integrate protein sequence similarity and protein-protein interaction network information and combines the pretraining model to extract the semantic features of the protein sequence. In addition, we design an attention-based graph neural network model, which can effectively extract node-level features from heterogeneous networks and predict protein function by measuring the similarity between protein nodes and gene ontology term nodes. Comparative experiments on the human dataset show that HNetGO achieves state-of-the-art performance on cellular component and molecular function branches.

MutCov: A pipeline for evaluating the effect of mutations in spike protein on infectivity and antigenicity of SARS-CoV-2

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), causing an outbreak of coronavirus disease 2019 (COVID-19), is a major threat to public health worldwide. Previous studies have shown that the spike protein of SARS-CoV-2 determines viral infectivity and major antigenicity. However, the spike protein has been undergoing various mutations, which bring a great challenge to the prevention and treatment of COVID-19. Here we present the MutCov, a pipeline for evaluating the effect of mutations in spike protein on infectivity and antigenicity of SARS-CoV-2 by calculating the binding free energy between spike protein and angiotensin-converting enzyme 2 (ACE2) or neutralizing monoclonal antibody (mAb). The predicted infectivity and antigenicity were highly consistent with biologically experimental results, and demonstrated that the MutCov achieved good prediction performance. In conclusion, the MutCov is of high importance for systematically evaluating the effect of novel mutations and improving the prevention and treatment of COVID-19. The source code and installation instruction of MutCov are freely available at http://jianglab.org.cn/MutCov.

Prediction of Metal Ion Binding Sites of Transmembrane Proteins

The metal ion binding of transmembrane proteins (TMPs) plays a fundamental role in biological processes, pharmaceutics, and medicine, but it is hard to extract enough TMP structures in experimental techniques to discover their binding mechanism comprehensively. To predict the metal ion binding sites for TMPs on a large scale, we present a simple and effective two-stage prediction method TMP-MIBS, to identify the corresponding binding residues using TMP sequences. At present, there is no specific research on the metal ion binding prediction of TMPs. Thereby, we compared our model with the published tools which do not distinguish TMPs from water-soluble proteins. The results in the independent verification dataset show that TMP-MIBS has superior performance. This paper explores the interaction mechanism between TMPs and metal ions, which is helpful to understand the structure and function of TMPs and is of great significance to further construct transport mechanisms and identify potential drug targets.

SYNBIP: synthetic binding proteins for research, diagnosis and therapy

The success of protein engineering and design has extensively expanded the protein space, which presents a promising strategy for creating next-generation proteins of diverse functions. Among these proteins, the synthetic binding proteins (SBPs) are smaller, more stable, less immunogenic, and better of tissue penetration than others, which make the SBP-related data attracting extensive interest from worldwide scientists. However, no database has been developed to systematically provide the valuable information of SBPs yet. In this study, a database named ‘Synthetic Binding Proteins for Research, Diagnosis, and Therapy (SYNBIP)’ was thus introduced. This database is unique in (a) comprehensively describing thousands of SBPs from the perspectives of scaffolds, biophysical & functional properties, etc.; (b) panoramically illustrating the binding targets & the broad application of each SBP and (c) enabling a similarity search against the sequences of all SBPs and their binding targets. Since SBP is a human-made protein that has not been found in nature, the discovery of novel SBPs relied heavily on experimental protein engineering and could be greatly facilitated by in-silico studies (such as AI and computational modeling). Thus, the data provided in SYNBIP could lay a solid foundation for the future development of novel SBPs. The SYNBIP is accessible without login requirement at both official (https://idrblab.org/synbip/) and mirror (http://synbip.idrblab.net/) sites.

ProtRe-CN: Protein Remote Homology Detection by Combining Classification Methods and Network Methods via Learning to Rank

Protein remote homology detection is one of fundamental research tasks for downstream analysis (i.e., protein structure and function prediction). Many advanced methods are proposed from different views with complementary detection ability, such as the classification method, the network method, and the ranking method. A framework integrating these heterogeneous methods is urgently desired to reduce the false positive rate and predictive bias. We propose a novel ranking method called ProtRe-CN by fusing the classification methods and network methods via Learning to Rank. Experimental results on the benchmark dataset and the independent dataset show that ProtRe-CN outperforms other existing state-of-the-art predictors. ProtRe-CN improves the detective performance via correcting the false positives in the ranking list by combining the heterogeneous methods. The web server of ProtRe-CN can be accessed at http://bliulab.net/ProtRe-CN.

A Brief History of De Novo Protein Design: Minimal, Rational, and Computational

Protein design has come of age, but how will it mature? In the 1980s and the 1990s, the primary motivation for de novo protein design was to test our understanding of the informational aspect of the protein-folding problem; i.e., how does protein sequence determine protein structure and function? This necessitated minimal and rational design approaches whereby the placement of each residue in a design was reasoned using chemical principles and/or biochemical knowledge. At that time, though with some notable exceptions, the use of computers to aid design was not widespread. Over the past two decades, the tables have turned and computational protein design is firmly established. Here, I illustrate this progress through a timeline of de novo protein structures that have been solved to atomic resolution and deposited in the Protein Data Bank. From this, it is clear that the impact of rational and computational design has been considerable: More-complex and more-sophisticated designs are being targeted with many being resolved to atomic resolution. Furthermore, our ability to generate and manipulate synthetic proteins has advanced to a point where they are providing realistic alternatives to natural protein functions for applications both in vitro and in cells. Also, and increasingly, computational protein design is becoming accessible to non-specialists. This all begs the questions: Is there still a place for minimal and rational design approaches? And, what challenges lie ahead for the burgeoning field of de novo protein design as a whole?

Synchrotron Radiation as a Tool for Macromolecular X-Ray Crystallography: a XXI Century Perspective

Intense X-rays available at powerful synchrotron beamlines provide macromolecular crystallographers with an incomparable tool for investigating biological phenomena on an atomic scale. The resulting insights into the mechanism's underlying biological processes have played an essential role and shaped biomedical sciences during the last 30 years, considered the "golden age" of structural biology. In this review, we analyze selected aspects of the impact of synchrotron radiation on structural biology. Synchrotron beamlines have been used to determine over 70% of all macromolecular structures deposited into the Protein Data Bank (PDB). These structures were deposited by over 13,000 different research groups. Interestingly, despite the impressive advances in synchrotron technologies, the median resolution of macromolecular structures determined using synchrotrons has remained constant throughout the last 30 years, at about 2 Å. Similarly, the median times from the data collection to the deposition and release have not changed significantly. We describe challenges to reproducibility related to recording all relevant data and metadata during the synchrotron experiments, including diffraction images. Finally, we discuss some of the recent opinions suggesting a diminishing importance of X-ray crystallography due to impressive advances in Cryo-EM and theoretical modeling. We believe that synchrotrons of the future will increasingly evolve towards a life science center model, where X-ray crystallography, Cryo-EM, and other experimental and computational resources and knowledge are encompassed within a versatile research facility. The recent response of crystallographers to the COVID-19 pandemic suggests that X-ray crystallography conducted at synchrotron beamlines will continue to play an essential role in structural biology and drug discovery for years to come.