Keywords: AlphaFold, machine learning, structural bioinformatics, structure prediction

Table 1.

Databases development and integrations after AlphaFold. The table organizes into categories the development of new databases and the integration of existing ones, following the release of the AlphaFold Database (AFDB).

AlphaFold integration in new databases

The establishment of a database derived from AF, that is AFDB, has significantly enhanced the efficiency of information sharing, emerging as an essential component for methodologies that depend on structural databases [51]. The AFDB contains predictions for all protein sequences annotated in the UniProt reference proteome, of length between 6 and 2700 amino acids, with the maximum limit decreasing to 1280 for noncurated sequences. A minimum length is required for having an informative MSA, while the maximum limits are set because of computational capacity. The integration of AFDB into pre‐existing databases represents a natural progression in response to AFDB's utility and relevance in the field. Indeed, key protein family databases such as InterPro and Pfam, as many others in the field [52, 53, 54, 55], have introduced dedicated pages for visualizing AF models. Additionally, the 2023 version of Ensembl [56] has incorporated AF models' visualization to map variant predictor effects onto the structure. Platforms such as PDB [8] and UniProt [57] are revolutionizing the management of predicted protein models within widely acknowledged databases, now incorporating AF predictions alongside experimental ones. The escalating volume of structures managed by these platforms has intensified the demand for scalability, a challenge efficiently met by renowned conformational search protocols, FoldSeek [42] and DALI [43], as well as newly developed ones [58]. Finally, new databases have been developed conveying newly available information to specific topics, for example, transmembrane proteins [33, 34]. It is important to note that this list is not exhaustive, and the landscape of AF integration is constantly evolving. Researchers are continuously finding new ways to incorporate AF into existing databases and cater to specific needs. These may include specialized databases for protein families, disease‐related proteins, or other specific fields. For instance, specialized protocols have emerged for analyzing kinases [59] and intrinsically disordered proteins [60]. The availability of such comprehensive data has significantly advanced high‐throughput and ‐omics research and has facilitated benchmarks of existing protocols, providing insights into their performance with predicted models [61, 62].

AlphaFold for experimental structure determination

The advent of AF, with its remarkable accuracy in predicting protein structures, has ushered in a transformative era for experimental structural biology, opening new avenues for investigating complex biological systems. One of the most significant contributions of AF lies in its application to Molecular Replacement (MR) in X‐ray crystallography. Traditionally, MR has relied on experimentally determined structures from the PDB, posing limitations when suitable homologs were not available. However, AF has revolutionized MR by offering high‐quality predicted protein structures as alternative search models [63, 64]. This breakthrough has significantly expanded the scope of MR, making it applicable to a broader range of proteins, including those with no known homologs in the PDB [65]. The integration of AF predictions into established software [36, 37, 38, 46, 47] underscores its rapid adoption and widespread impact across various macromolecular structure determination methodologies. In PHENIX, AF models can be utilized through a dedicated set of functions [38, 46], including a PHENIX‐AF webservice to run predictions remotely from the GUI [47], their import from ColabFold [66], their trimming and splitting into single domains, and finally their positioning in unit cells. The resulting models can be examined with PHENIX validation tools to identify and manually fix any problematic areas. Similarly, CCP4 provides seamless integration of AF models for MR [35, 39, 41], interacting with the AFDB.

AF has been also integrated into various Cryo‐electron microscopy (cryo‐EM) pipelines, streamlining the workflow and improving both speed and accuracy [67]. Traditionally, building atomic models from cryo‐EM density maps has been a laborious and error‐prone process. AF predictions provide researchers with a high‐quality starting point. Software tools such as MrParse [36] and UCSF ChimeraX [40] can seamlessly access the AFDB, allowing researchers to achieve precise protein positioning within the cryo‐EM map by superimposing the AF model, which significantly improves the final model's quality, while substantially reducing manual building time. ISOLDE is another tool incorporating AF models during the refinement process [38]. This allows ISOLDE to utilize the predicted information alongside the cryo‐EM data, potentially leading to a more refined and accurate final structure in agreement with experimental data.

In particular, AF predictions can significantly improve the quality of cryo‐EM reconstructions, especially when dealing with data with low resolution and can be used as accurate starting models to fit components into cryo‐EM densities [67, 68, 69, 70, 71]. This is particularly helpful for determining the structures of large protein assemblies, such as the nucleopore complex [72]. Here, the authors utilized AF to enhance the structural determination of the nuclear pore complex's (NPC) cytoplasmic ring using integrative cryo‐EM. The high‐accuracy predictions of AF were crucial in providing detailed atomic models, accurately positioning proteins within the cryo‐EM density maps, and bridging gaps in incomplete experimental data. This integrative approach led to a more comprehensive and accurate model, revealing intricate protein interactions and conformations.

In another recent study, researchers working to solve the structure of the mycobacterial lipid transporter Mce1 were able to assign density to a previously unknown subunit of the complex, LucB protein. They were able to perform a structural search of a density‐derived poly‐Ala model against a large number of predictions in AFDB, which returned LucB as a hit. The assignment was subsequently experimentally validated [73].

Unleashing the potential of AI in protein structure prediction

As the creation of AFDB streamlined the integration of AF into commonly utilized databases and tools [44, 45, 48, 49, 50], in a similar vein, releasing the source code of AF fostered the development of other AI‐based tools for protein structure prediction (Table 2). Consequently, AF promoted the idea that, as in other fields, AI was progressing exponentially, and promising results could also be achieved in structural biology. Although neural networks have long been applied to structural prediction, the point at which AI‐based predictions began to significantly outperform traditional methods coincided with the introduction of Transformer models [88], exemplified by the model used in AF.

Almost in parallel with AF, the RosettaFold [84] neural network for protein structure prediction has emerged. The “two‐track” version of the network was outperforming trRosetta [2]. However, a performance improvement, approaching that of AF, has been observed with the development of a “three‐track” neural network. This latter was inspired by the features that contributed to the performance of AF, reworking them to operate in 3D coordinate space in order to establish a closer relationship between sequence, residue–residue distances and orientations, and atomic coordinates.

Methods such as AF, which rely on co‐evolution information extracted from MSA, inevitably hinder the possibility of prediction when an accurate MSA is lacking, as is the case of orphan proteins. The optimization and broader application of language models in different areas have paved a new direction in protein modeling too. These advancements have led to the development of MSA‐free approaches that are both computationally efficient and highly accurate. By leveraging the contextual understanding of protein sequences provided by language models, these tools can generate accurate predictions even in the absence of extensive sequence homology. Notable examples of these advancements include ESMFold [31], OmegaFold [83], and AminoBERT [89]. A significant achievement has been made with the large‐scale application of ESMFold, whose training data retrieval has been inspired by AF, culminating in the creation of their Metagenomic Atlas, comprising over 700 million predicted structures.

Protein ligands and cofactors

The precise prediction of protein–ligand complexes plays a crucial role in overcoming various challenges in targeted therapies. As anticipated, after CASP14, several AI‐based methods for the universal modeling of proteins in combination with small molecules have been developed [87, 92]. While the performance of such methods is still not outperforming classical, physico‐chemical‐based approaches, it is expected that the parallel growth of both protein structure prediction and protein–ligand docking will lead to mutual benefits in the accurate prediction of protein–ligand complexes.

The inability to predict proteins in complex with ligands and cofactors, which is one of the main pinpointed limitations of AF, has been partially addressed by AlphaFill [32], which builds on AFDB. With this protocol, AF models can be enriched by tapping into the extensive resources of the PDB‐REDO [93] and CoFactor databases [94]. The protocol, validated for correlation in terms of root mean square deviation (RMSD) with the experimental structure both globally and locally, involves transplanting the most common ligands and cofactors from a sequence homologous to the AF model. This is achieved through a local structural superimposition of the model and the homologous experimental structure.

Multidomain modeling

Homology modeling tools have emerged as an effective complement in scenarios where AF may not provide complete solutions. Structure prediction of GPCRs make up a paradigmatic case for the quality assessment of the predictions for a specific protein family, for which the structure determination has been critical so far. A comparative study involving AF, RoseTTAFold and MODELLER [95], confirmed the expected higher performance of MODELLER, whenever a high‐quality structural homolog is used as a template. Since this higher performance is evident when assessing interdomain positioning, a combined approach, utilizing both template‐based and template‐free methods, can yield effective results. Indeed, on this concept, recent pipelines have been developed to leverage the strengths of both approaches. One such example is AlphaMod [96], an automated pipeline that fuses AF with MODELLER, a well‐established template‐based modeling software. In a similar fashion, MoDAFold [97], combines AF with MD simulations to predict the structure of missense proteins with higher accuracy.

To address the problem of ab‐initio multidomain protein modeling, DeepAssembly, a new computational protocol for assembling multidomain proteins and complexes, was recently developed [77]. DeepAssembly uses a deep learning network to predict interdomain interactions, and then employs a population‐based evolutionary algorithm to assemble domains into complete structures. This approach outperforms AF in predicting interdomain distances in multidomain proteins and improves accuracy for low‐confidence structures in the AFDB.

Protein–protein complexes

The “Assembly” category in CASP competitions has shown a notable upward trajectory since its introduction in CASP12 [98]. Even though it saw limited participation, it presented a significant opportunity for progress and quickly captured the interest of the scientific community, witnessing a surge in engagement in the subsequent years [99], and has now become one of the most hyped categories. Notably, Deepmind group did not take part in this competition, as their AF multimer version was not competitive enough [100]. However, building on the limits of AF, the quality of predictions within this category had improved substantially in CASP15 [101], with the first‐ranked DMFold‐Multimer [78], heavily influenced by the AF structural module. AlphaFold‐predicted pairwise subunit interactions can also be exploited for assembly prediction, as shown in new advancements that focused on MSA sampling, for example, AFsample and MULTICOM [74, 76, 102].

PROTAC modeling can be considered another pertinent example of the importance of protein complex prediction, even for nonphysiological interactions, in which AF fails to obtain accurate predictions [103]. Historically centered around a limited set of E3 ligases, the field is witnessing a shift with the discovery of new E3 ligase structures, opening avenues for their utilization and the rational design of ligands [104]. Yet, a notable gap remains in predicting the proper orientation of the E3–ligase complex concerning the target protein, which is essential for establishing a solid foundation for the design of ligands and linkers. Thus far, addressing such a challenge has involved utilizing a combination of tools for protein structure prediction, such as RosettaFold, along with protein docking techniques [105]. However the shift toward universal modeling is expected to provide a method to consider the orientation of the E3 Ligase–Target protein complex and the design of an appropriate bivalent ligand, at once.

Protein–nucleic acids complexes

Understanding protein–nucleic acid interactions is particularly vital for decoding complex biological processes such as gene expression and genome repair, but accurate prediction presents significant computational hurdles due to the complexity of biomolecular interactions [106]. This knowledge is also pivotal for precision applications in genome editing, such as the engineering of Cas proteins, which are integral to technologies such as CRISPR‐Cas editing [107].

On the wave of universal modeling, the concepts and techniques underlying AF and RoseTTAFold have been extended to the prediction of the structures of nucleic acids and protein–nucleic acid complexes, leading to the development of AlphaFold3 (AF3) [75] and RosettaFoldNA [86], followed by RoseTTAFold All‐Atom [85]. Based on the RosettaFold three‐track neural networks for molecule representation, RoseTTAFold All‐Atom enhances this framework by incorporating atomic‐level details and chemical representations across various dimensions into a diffusion model. As input representation, the first track of RoseTTAFold All‐Atom encodes the sequence information of proteins and nucleic acids, including amino acid types and nucleotide bases. For nonpolymer atoms, it encodes their chemical element type. The second track represents pairwise information between atoms, including chemical bonds and distances. The last track includes the 3D coordinates of atoms or residues, along with information about chirality. The network of RoseTTAFold All‐Atom employs attention mechanisms to weigh the importance of different input features, allowing for dynamic and context‐dependent learning, and iteratively refines the predicted structure by updating the 3D coordinates based on the information from all three tracks. This integration significantly boosts the resolution and accuracy of the predicted molecular structures.

The AF3 architecture builds upon its predecessor by incorporating diffusion models as a generative module specifically designed for 3D structure generation [75]. Input data need additional preprocessing given the different molecular types the model has to handle. To do so, the raw inputs, the MSA, and the ligand conformers are converted into three different embeddings, namely the “Input,” “Pair,” and “Single” representations. The “Input” representation includes basic atomic and residue information such as type, position, and charge. The “Single” representation groups atoms by their amino acids or nucleotides, adding contextual information. The “Pair” representation captures spatial relationships in protein and DNA/RNA sequences, enriched with template and co‐evolutionary data. The “Pairformer,” which receives the enriched “Pair” representation, refines single and pair representations through recycling steps to produce a structural hypothesis, which then conditions the diffusion module for generating 3D coordinates. Despite the significance of AF's universal modeling generalization, the release of AF3 was notably different. The absence of source code and server constraints make it difficult to understand how AF3 generalizes and prevent its verification and replication.

Of note, such advancements boosted the research toward the application of deep learning methods also for predicting the secondary structures of DNA and RNA alone [108, 109, 110, 111].

Protein dynamics

Exploring the dynamics of proteins has consistently presented a complex challenge, not just within the realm of computational predictions but also in experimental approaches. This challenge is linked to the prediction of protein complexes, as the interaction with other molecules often induces significant conformational changes in proteins [112]. The issue has gained renewed attention with the advent of AF, which, despite its advancements, tends to favor predictions biased toward more commonly represented conformations in its training data [113, 114]. In scenarios where conformational changes are subtle, implementing postprediction processing techniques and enhancing the accuracy of preliminary models (i.e., “refinement”) emerges as a viable strategy to mitigate this limitation. CASP10 assessment of the refinement category [115] has highlighted the effectiveness of refinement methods, especially those employing molecular dynamics, in producing conformations that in their highest accuracy find also suited application in MR.

In the era following CASP‐15, the endeavor to encompass the vast diversity of protein structural conformations can be expanded to most difficult and generalizable tasks. The approaches proposed leverage on tuning the MSA, that is, masking some positions, to guide the AF algorithm toward various conformational states [116, 117, 118, 119, 120]. The adoption of the flow‐matching method has significantly enhanced the accuracy of predicting protein conformational ensembles. This approach involves training generative models to closely replicate the distribution of protein conformations observed in experimental data or simulations [121, 122]. Another recent study showed that machine learning models can be trained on simulation data to directly create realistic protein structures without the need for extensive sampling, which significantly reduces computational cost [123]. The authors demonstrated this with a model called idpGAN, trained on coarse‐grained simulations of intrinsically disordered peptides. The model can predict new structures for sequences it has not seen before, proving its ability to generalize beyond the training data.

This evolution in strategy underscores the ongoing effort to refine computational tools for protein structure prediction, aiming for a more comprehensive and accurate representation of protein behavior.

Orphan proteins

Methods such as AF, which rely on co‐evolution information extracted from Multiple Sequence Alignments (MSA), face challenges when predicting the structure of orphan proteins that lack accurate MSAs. This limitation has spurred the development of MSA‐free approaches that offer computationally efficient and highly accurate alternatives for predicting orphan protein structures. Notable examples of these advancements include ESM‐Fold [31] and OmegaFold [83], which leverage language models to bypass the limitations of MSA dependence. These models have shown promising results in predicting orphan protein structures, providing a valuable tool for understanding these previously enigmatic proteins.

A significant breakthrough in this field has been achieved with the large‐scale application of ESMFold, whose training data retrieval was inspired by AF. This effort culminated in the creation of their Metagenomic Atlas, a comprehensive repository of over 700 million predicted structures, including numerous orphan proteins. This atlas represents a major step forward in our understanding of the vast and diverse world of orphan proteins, offering valuable insights into their structures and potential functions.

Furthermore, the success of ESMFold and OmegaFold in predicting orphan protein structures has paved the way for further research and development in this area. Ongoing efforts are focused on refining these models, exploring novel MSA‐free approaches, and expanding the Metagenomic Atlas to include an even wider range of orphan proteins. Notable examples of these approaches are represented by HelixFold‐Single [81] and RGN [89]. The ultimate goal is to develop robust and reliable tools that can accurately predict the structures of all orphan proteins, unlocking their structural/functional peculiarities and contributing to our understanding of the complex biological processes they are involved in [124].

De novo protein design

With 20 naturally occurring amino acids, a protein consisting of 100 amino acids could theoretically manifest in 20^100 different sequence variations. Given the diversity of protein sizes, the theoretical number of possible proteins far exceeds the number of proteins identified by nature [129]. De novo protein design leverages computational algorithms and biophysical principles to engineer novel proteins with tailored functions that, in billions of years of tinkering, Nature has never produced. Computational tools, often grounded in physics‐based energy functions and machine learning models, enable the exploration of vast sequence spaces and the prediction of protein structures. Recently, several protein design tools have been introduced by researchers (Table 3), which showcase substantial progress in the field. For a detailed review of de novo protein design, see Ref. [145]. Here, we will focus on two very recent and state‐of‐the‐art advancements in all‐atoms approaches, that is, RFdiffusion All‐Atom [85, 144] and ESM3 [135].

Recently, diffusion models have started to be used in protein and peptide design [146, 147]. These models, originally popularized in the field of generative art for their ability to create detailed and high‐fidelity images [148, 149], offer several advantages that make them suitable for protein modeling. Diffusion models work by iteratively refining an initial random structure into a coherent final one through a process that simulates the gradual “denoising” of a protein's atomic coordinates [150]. This iterative refinement allows for capturing the nuanced details of protein structures that are crucial for understanding their functions and interactions. While it is generally true that most neural networks are adept at capturing the global properties of protein structures [151], diffusion models offer an edge in generating high‐diversity folds, which in turn can be conditioned through a wide variety of inputs or design objectives [136]. In this field, the enhanced RFdiffusion All‐Atom model incorporates diverse biological building blocks such as DNA, RNA, ions, and small molecules, expanding the scope of protein design possibilities. When coupled to other tools such as ProteinMPNN [152], LigandMPNN [153], and AF (see, e.g., a design pipeline of heme‐binding proteins, available at: https://github.com/ikalvet/heme_binder_diffusion), it opens avenues for designing proteins with unprecedented sequences, structures, and functions. ESM3 [135] is another cutting‐edge AI model that can understand and design protein sequences, structures, and functions, using a frontier multimodal generative language model. The latter has been trained on a massive dataset of entries and can be prompted with any combination of sequence, structure, or function information to generate new proteins. Notably, it can generate proteins with characteristics not seen in nature, demonstrating its creativity in problem‐solving. As an example of its capabilities, ESM3 generated a new green fluorescent protein (esmGFP), which is significantly different from any known natural protein. This level of novelty is comparable to the amount of change that occurs in natural proteins over hundreds of millions of years of evolution. This demonstrates the potential of ESM3 as a powerful tool in protein engineering.

These techniques are revolutionizing protein engineering by enabling the rapid design of novel proteins with desired properties and, most importantly, have been experimentally validated, which in the end serves as the ultimate benchmark for the efficacy of predictive tools.

Antibody design

Structure prediction of antibodies could be considered as a specialized area of protein structure prediction. In developing therapeutic antibodies, vaccines, and treatments for autoimmune disorders, the structural prediction of antibodies has historically depended on homology modeling, due to their highly evolutionarily conserved Y‐shaped scaffold [102]. However, predicting antibody structures, especially the highly variable complementary determining regions (CDR)‐H3 loop [154], remains challenging. Several methods have been developed to address this, utilizing both ab initio protocols and machine learning techniques. Ab initio protocols such as OptCDR [155], RosettaAntibody [156], and AbDesign [157] tackle this problem by redesigning CDRs to enhance antibody stability and affinity by optimizing conformational and free energy changes in specific residues. RosettaAntibody, for example, can perform de novo antibody design or affinity maturation of existing antibodies by classifying the antibody into regions, including the framework, canonical loops, and HCDR3 loop.

For more complex tasks, innovative antibody design protocols have emerged, including DeepH3, DeepAb, IgFold, and ABlooper, showcasing the power of machine learning techniques in antibody structural predictions [130, 132, 133, 158]. DeepH3, a deep residual neural network, identifies near‐native CDR‐H3 loops and improves the average RMSD of prediction compared to the standard Rosetta energy function. AbLooper employs Equivariant Graph Neural Networks to predict CDR structures, producing accurate antibody models efficiently.

Furthermore, pretrained language models have proven effective in inferring full atomic‐level protein structures. DeepAb leverages an antibody pretrained language model with recurrent neural network to reconstruct the entire antibody variable region, generating more precise structures compared with alternatives. IgFold, inspired by DeepAb, utilizes a pretrained language model trained on natural antibody sequences and graph networks to directly predict backbone atom coordinates, offering high speed, accuracy, and nanobody modeling capabilities.

While the remarkable advancements in domain prediction contribute to highly accurate models of immunoglobulin domains or target epitopes, the precise orientation and interaction between the CDR loops and the target epitope are areas that require further refinement [159, 160, 161]. These aspects of antibody–antigen interaction are more likely to benefit from the latest breakthroughs in predicting multimeric complexes, highlighting a crucial direction for future advancements in antibody design [130, 162, 163]. Sculptor is a new algorithm that addresses this challenge using deep generative design to create antibodies that bind to specific epitopes [164]. It does this by jointly searching for the best positions, interactions, and shapes of the protein scaffold. It then designs a protein backbone that complements the target.

In summary, while accurately predicting antibody structures, especially CDR‐H3, remains a challenge, significant progress has been made through the development of various computational methods and the integration of machine learning techniques and pretrained language models. These advancements hold promise for accelerating antibody design and engineering efforts, ultimately contributing to the development of more effective therapeutic antibodies.

Peptide design

The development of therapeutic peptides hinges on the ability to design peptidic binders that target specific proteins of interest. Traditionally, peptide design has relied on a combination of rational design, simulation, and screening techniques [165]. Similarly, early AI‐based approaches to peptide design, which were adapted from protein design methods, employed various techniques such as inverse design (e.g., ProteinMPNN [152]), peptide‐specific methods (e.g., PepMLM [166]), and generative models (e.g., MaSIF‐Seed [143]). These approaches all aimed to design new peptide binders starting from the target protein. By leveraging different architectures, it is now possible to focus on the de‐novo design of peptide sequences. These approaches are particularly valuable for their ability to capture the distribution of amino acids that confer a set of functionalities and activities, such as antimicrobial, anticancer, immunogenic properties, or signal peptide functions. Variational autoencoders (VAEs), generative adversarial networks (GANs), and diffusion models have emerged as viable options. As an example, a recent multistep sequence generation algorithm was proposed [140]. The deep learning‐based generative model Gated Recurrent Unit based variational autoencoder (GRU‐based VAE) and the Metropolis Hasting (MH) sampling algorithm efficiently generate new peptide sequences. The binding affinity of generated peptides is then evaluated using physics‐based methods, such as molecular dynamics (MD) simulations. Several GANs have been trained for peptide design, tailored to specific use cases such as immunogenic, antimicrobial [167, 168, 169], and antiviral peptides [170]. Other examples of GANs, such as HelixGAN [142], have been specifically trained to focus on the design of helical peptides, and similarly, a diffusion model called HelixDiff [141] has been developed with the same objective. In a recent work [137], a diffusion‐based model (FoldingDiff) generating high‐quality backbone peptides (up to 128 residues) via a procedure inspired by the natural folding process, is presented. FoldingDiff uses a sequence of dihedral angles capturing the relative orientation of the constituent backbone atoms and generates stable folded peptides by denoising from an unfolded structure. Moreover, the development of combined and fine‐tuned approaches is becoming increasingly common. Latent space diffusion models, in particular, are gaining traction, and AMP‐Diffusion [171] exemplifies their application by harnessing the power of latent representations and the flexibility of diffusion processes to enhance the generation of antimicrobial peptides.

• 1. Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, Tunyasuvunakool K, Bates R, Žídek A, Potapenko A et al. (2021) Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 2. Pereira J, Simpkin AJ, Hartmann MD, Rigden DJ, Keegan RM and Lupas AN (2021) High‐accuracy protein structure prediction in CASP14. Proteins 89, 1687–1699. [DOI] [PubMed] [Google Scholar]
• 3. Varadi M, Anyango S, Deshpande M, Nair S, Natassia C, Yordanova G, Yuan D, Stroe O, Wood G, Laydon A et al. (2022) AlphaFold protein structure database: massively expanding the structural coverage of protein‐sequence space with high‐accuracy models. Nucleic Acids Res 50, 439–444. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 4. Varadi M, Bertoni D, Magana P, Paramval U, Pidruchna I, Radhakrishnan M, Tsenkov M, Nair S, Mirdita M, Yeo J et al. (2024) AlphaFold protein structure database in 2024: providing structure coverage for over 214 million protein sequences. Nucleic Acids Res 52, 368–375. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 5. Waterhouse A, Bertoni M, Bienert S, Studer G, Tauriello G, Gumienny R, Heer FT, de Beer TAP, Rempfer C, Bordoli L et al. (2018) SWISS‐MODEL: homology modelling of protein structures and complexes. Nucleic Acids Res 46, 296–303. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 6. Porta‐Pardo E, Ruiz‐Serra V, Valentini S and Valencia A (2022) The structural coverage of the human proteome before and after AlphaFold. PLoS Comput Biol 18, e1009818. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 7. David A, Islam S, Tankhilevich E and Sternberg MJE (2022) The AlphaFold database of protein structures: a biologist's guide. J Mol Biol 434, 167336. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 8. Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, Shindyalov IN and Bourne PE (2000) The Protein Data Bank. Nucleic Acids Res 28, 235–242. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 9. Sippl MJ, Lackner P, Domingues FS and Koppensteiner WA (1999) An attempt to analyse progress in fold recognition from CASP1 to CASP3. Proteins Suppl 3, 226–230. [DOI] [PubMed] [Google Scholar]
• 10. Huang YJ, Zhang N, Bersch B, Fidelis K, Inouye M, Ishida Y, Kryshtafovych A, Kobayashi N, Kuroda Y, Liu G et al. (2021) Assessment of prediction methods for protein structures determined by NMR in CASP14: impact of AlphaFold2. Proteins 89, 1959–1976. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 11. Moult J (2005) A decade of CASP: progress, bottlenecks and prognosis in protein structure prediction. Curr Opin Struct Biol 15, 285–289. [DOI] [PubMed] [Google Scholar]
• 12. Johnson MS, Srinivasan N, Sowdhamini R and Blundell TL (1994) Knowledge‐based protein modelling. Crit Rev Biochem Mol Biol 29, 1–68. [DOI] [PubMed] [Google Scholar]
• 13. Rost B and Sander C (1996) Bridging the protein sequence–structure gap by structure predictions. Annu Rev Biophys Biomol Struct 25, 113–136. [DOI] [PubMed] [Google Scholar]
• 14. Sali A and Blundell TL (1993) Comparative protein modelling by satisfaction of spatial restraints. J Mol Biol 234, 779–815. [DOI] [PubMed] [Google Scholar]
• 15. Bonneau R, Tsai J, Ruczinski I, Chivian D, Rohl C, Strauss CE and Baker D (2001) Rosetta in CASP4: progress in ab initio protein structure prediction. Proteins Suppl 5, 119–126. [DOI] [PubMed] [Google Scholar]
• 16. Rohl CA, Strauss CE, Misura KM and Baker D (2004) Protein structure prediction using Rosetta. Methods Enzymol 383, 66–93. [DOI] [PubMed] [Google Scholar]
• 17. Zhang Y (2008) I‐TASSER server for protein 3D structure prediction. BMC Bioinformatics 9, 40. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 18. Zhang Y and Skolnick J (2004) Automated structure prediction of weakly homologous proteins on a genomic scale. Proc Natl Acad Sci USA 101, 7594–7599. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 19. Göbel U, Sander C, Schneider R and Valencia A (1994) Correlated mutations and residue contacts in proteins. Proteins 18, 309–317. [DOI] [PubMed] [Google Scholar]
• 20. Giraud BG, Heumann JM and Lapedes AS (1999) Superadditive correlation. Phys Rev 59, 4983–4991. [DOI] [PubMed] [Google Scholar]
• 21. Afonnikov DA, Kondrakhin YV, Titov II and Kolchanov NA (1997) Detecting direct correlation between positions in multiple alignment of amino‐acid sequences. In Computer Science and Biology. Genome Informatics: Function, Structure, Phylogeny (Frishman D and Mewes HW, eds), pp. 87–98. Proceedings of the German Conference on Bioinformatics. [Google Scholar]
• 22. Weigt M, White RA, Szurmant H, Hoch JA and Hwa T (2009) Identification of direct residue contacts in protein–protein interaction by message passing. Proc Natl Acad Sci USA 106, 67–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 23. Burger L and van Nimwegen E (2010) Disentangling direct from indirect co‐evolution of residues in protein alignments. PLoS Comput Biol 6, e1000633. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 24. Marks DS, Colwell LJ, Sheridan R, Hopf TA, Pagnani A, Zecchina R and Sander C (2011) Protein 3D structure computed from evolutionary sequence variation. PLoS One 6, e28766. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 25. Skwark MJ, Raimondi D, Michel M and Elofsson A (2014) Improved contact predictions using the recognition of protein like contact patterns. PLoS Comput Biol 10, e1003889. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 26. Wang S, Sun S, Li Z, Zhang R and Xu J (2017) Accurate de novo prediction of protein contact map by ultra‐deep learning model. PLoS Comput Biol 13, e1005324. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 27. Senior AW, Evans R, Jumper J, Kirkpatrick J, Sifre L, Green T, Qin C, Žídek A, Nelson AWR, Bridgland A et al. (2020) Improved protein structure prediction using potentials from deep learning. Nature 577, 706–710. [DOI] [PubMed] [Google Scholar]
• 28. Mirabello C and Wallner B (2019) rawMSA: end‐to‐end deep learning using raw multiple sequence alignments. PLoS One 14, e0220182. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 29. Kryshtafovych A, Antczak M, Szachniuk M, Zok T, Kretsch RC, Rangan R, Pham P, Das R, Robin X, Studer G et al. (2023) New prediction categories in CASP15. Proteins 91, 1550–1557. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 30. Kryshtafovych A, Schwede T, Topf M, Fidelis K and Moult J (2021) Critical assessment of methods of protein structure prediction (CASP)–round XIV. Proteins 89, 1607–1617. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 31. Lin Z, Akin H, Rao R, Hie B, Zhu Z, Lu W, Smetanin N, Verkuil R, Kabeli O, Shmueli Y et al. (2023) Evolutionary‐scale prediction of atomic‐level protein structure with a language model. Science 379, 1123–1130. [DOI] [PubMed] [Google Scholar]
• 32. Hekkelman ML, de Vries I, Joosten RP and Perrakis A (2023) AlphaFill: enriching AlphaFold models with ligands and cofactors. Nat Methods 20, 205–213. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 33. Dobson L, Szekeres LI, Gerdán C, Langó T, Zeke A and Tusnády GE (2023) TmAlphaFold database: membrane localization and evaluation of AlphaFold2 predicted alpha‐helical transmembrane protein structures. Nucleic Acids Res 51, D517–D522. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 34. Pei J and Cong Q (2023) AFTM: a database of transmembrane regions in the human proteome predicted by AlphaFold. Database (Oxford) 2023, baad008. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 35. Agirre J, Atanasova M, Bagdonas H, Ballard CB, Baslé A, Beilsten‐Edmands J, Borges RJ, Brown DG, Burgos‐Mármol JJ, Berrisford JM et al. (2023) The CCP4 suite: integrative software for macromolecular crystallography. Acta Crystallogr D 79, 449–461. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 36. Simpkin AJ, Thomas JMH, Keegan RM and Rigden DJ (2022) MrParse: finding homologues in the PDB and the EBI AlphaFold database for molecular replacement and more. Acta Crystallogr D 78, 553–559. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 37. Keegan RM and Winn MD (2008) MrBUMP: an automated pipeline for molecular replacement. Acta Crystallogr D 64, 119–124. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 38. Oeffner RD, Croll TI, Millán C, Poon BK, Schlicksup CJ, Read RJ and Terwilliger TC (2022) Putting AlphaFold models to work with phenix.process_predicted_model and ISOLDE. Acta Crystallogr D 78, 1303–1314. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 39. Simpkin AJ, Caballero I, McNicholas S, Stevenson K, Jiménez E, Sánchez Rodríguez F, Fando M, Uski V, Ballard C, Chojnowski G et al. (2023) Predicted models and CCP4. Acta Crystallogr D 79, 806–819. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 40. Pettersen EF, Goddard TD, Huang CC, Meng EC, Couch GS, Croll TI, Morris JH and Ferrin TE (2021) UCSF ChimeraX: structure visualization for researchers, educators, and developers. Protein Sci 30, 70–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 41. Emsley P and Cowtan K (2004) Coot: model‐building tools for molecular graphics. Acta Crystallogr D 60, 2126–2132. [DOI] [PubMed] [Google Scholar]
• 42. van Kempen M, Kim SS, Tumescheit C, Mirdita M, Lee J, Gilchrist CLM, Söding J and Steinegger M (2024) Fast and accurate protein structure search with Foldseek. Nat Biotechnol 42, 243–246. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 43. Holm L (2022) Dali server: structural unification of protein families. Nucleic Acids Res 50, W210–W215. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 44. Waterhouse AM, Procter JB, Martin DMA, Clamp M and Barton GJ (2009) Jalview version 2 – a multiple sequence alignment editor and analysis workbench. Bioinformatics 25, 1189–1191. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 45. Sehnal D, Bittrich S, Deshpande M, Svobodová R, Berka K, Bazgier V, Velankar S, Burley SK, Koča J and Rose AS (2021) Mol* Viewer: modern web app for 3D visualization and analysis of large biomolecular structures. Nucleic Acids Res 49, W431–W437. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 46. Afonine PV, Poon BK, Read RJ, Sobolev OV, Terwilliger TC, Urzhumtsev A and Adams PD (2018) Real‐space refinement in PHENIX for cryo‐EM and crystallography. Acta Crystallogr D 74, 531–544. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 47. Poon BK, Terwilliger TC and Adams PD (2024) The Phenix‐AlphaFold webservice: enabling AlphaFold predictions for use in Phenix. Protein Sci 33, e4992. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 48. Pfab J, Phan NM and Si D (2021) DeepTracer for fast de novo cryo‐EM protein structure modeling and special studies on CoV‐related complexes. Proc Natl Acad Sci USA 118, e2017525118. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 49. Giri N and Cheng J (2023) Improving protein–ligand interaction modeling with cryo‐EM data, templates, and deep learning in 2021 ligand model challenge. Biomolecules 13, 132. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 50. He J, Lin P, Chen J, Cao H and Huang SY (2022) Model building of protein complexes from intermediate‐resolution cryo‐EM maps with deep learning‐guided automatic assembly. Nat Commun 13, 4066. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 51. Varadi M and Velankar S (2023) The impact of AlphaFold protein structure database on the fields of life sciences. Proteomics 23, e2200128. [DOI] [PubMed] [Google Scholar]
• 52. Paysan‐Lafosse T, Blum M, Chuguransky S, Grego T, Pinto BL, Salazar GA, Bileschi ML, Bork P, Bridge A, Colwell L et al. (2023) InterPro in 2022. Nucleic Acids Res 51, 418–427. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 53. Mistry J, Chuguransky S, Williams L, Qureshi M, Salazar GA, Sonnhammer ELL, Tosatto SCE, Paladin L, Raj S, Richardson LJ et al. (2021) Pfam: the protein families database in 2021. Nucleic Acids Res 49, 412–419. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 54. Bairoch A and Apweiler R (2000) The SWISS‐PROT protein sequence database and its supplement TrEMBL in 2000. Nucleic Acids Res 28, 45–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 55. Szklarczyk D, Franceschini A, Wyder S, Forslund K, Heller D, Huerta‐Cepas J, Simonovic M, Roth A, Santos A, Tsafou KP et al. (2015) STRING v10: protein–protein interaction networks, integrated over the tree of life. Nucleic Acids Res 43, 447–452. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 56. Martin FJ, Amode MR, Aneja A, Austine‐Orimoloye O, Azov AG, Barnes I, Becker A, Bennett R, Berry A, Bhai J et al. (2023) Ensembl 2023. Nucleic Acids Res 51, 933–941. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 57. UniProt Consortium (2023) UniProt: the universal protein knowledgebase in 2023. Nucleic Acids Res 51, 523–531. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 58. Aderinwale T, Bharadwaj V, Christoffer C, Terashi G, Zhang Z, Jahandideh R, Kagaya Y and Kihara D (2022) Real‐time structure search and structure classification for AlphaFold protein models. Commun Biol 5, 316. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 59. Faezov B and Dunbrack RL Jr (2023) AlphaFold2 models of the active form of all 437 catalytically competent human protein kinase domains. bioRxiv. doi: 10.1101/2023.07.21.550125 [DOI] [Google Scholar]
• 60. Piovesan D, Del Conte A, Clementel D, Monzon AM, Bevilacqua M, Aspromonte MC, Iserte JA, Orti FE, Marino‐Buslje C and Tosatto SCE (2023) MobiDB: 10 years of intrinsically disordered proteins. Nucleic Acids Res 51, 438–444. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 61. Jakubec D, Skoda P, Krivak R, Novotny M and Hoksza D (2022) PrankWeb 3: accelerated ligand‐binding site predictions for experimental and modelled protein structures. Nucleic Acids Res 50, 593–597. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 62. Saito M, Xu P, Faure G, Maguire S, Kannan S, Altae‐Tran H, Vo S, Desimone A, Macrae RK and Zhang F (2023) Fanzor is a eukaryotic programmable RNA‐guided endonuclease. Nature 620, 660–668. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 63. Terwilliger TC, Afonine PV, Liebschner D, Croll TI, McCoy AJ, Oeffner RD, Williams CJ, Poon BK, Richardson JS, Read RJ et al. (2023) Accelerating crystal structure determination with iterative AlphaFold prediction. Acta Crystallogr D 79, 234–244. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 64. McCoy AJ, Sammito MD and Read RJ (2022) Implications of AlphaFold2 for crystallographic phasing by molecular replacement. Acta Crystallogr D 78, 1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 65. Millán C, Keegan RM, Pereira J, Sammito MD, Simpkin AJ, McCoy AJ, Lupas AN, Hartmann MD, Rigden DJ and Read RJ (2021) Assessing the utility of CASP14 models for molecular replacement. Proteins 89, 1752–1769. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 66. Mirdita M, Schütze K, Moriwaki Y, Heo L, Ovchinnikov S and Steinegger M (2022) ColabFold: making protein folding accessible to all. Nat Methods 19, 679–682. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 67. Rantos V, Karius K and Kosinski J (2022) Integrative structural modeling of macromolecular complexes using Assembline. Nat Protoc 17, 152–176. [DOI] [PubMed] [Google Scholar]
• 68. Alshammari M, Wriggers W, Sun J and He J (2022) Refinement of AlphaFold2 models against experimental and hybrid cryo‐EM density maps. QRB Discov 3, e16. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 69. DiIorio MC and Kulczyk AW (2023) Novel artificial intelligence‐based approaches for ab initio structure determination and atomic model building for cryo‐electron microscopy. Micromachines (Basel) 14, 1674. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 70. Ziegler SJ, Mallinson SJB, St. John PC and Bomble YJ (2021) Advances in integrative structural biology: towards understanding protein complexes in their cellular context. Comput Struct Biotechnol J 19, 214–225. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 71. Terwilliger TC, Poon BK, Afonine PV, Schlicksup CJ, Croll TI, Millán C, Richardson JS, Read RJ and Adams PD (2022) Improved AlphaFold modeling with implicit experimental information. Nat Methods 19, 1376–1382. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 72. Fontana P, Dong Y, Pi X, Tong AB, Hecksel CW, Wang L, Fu TM, Bustamante C and Wu H (2022) Structure of cytoplasmic ring of nuclear pore complex by integrative cryo‐EM and AlphaFold. Science 376, eabm9326. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 73. Chen J, Fruhauf A, Fan C, Ponce J, Ueberheide B, Bhabha G and Ekiert DC (2023) Structure of an endogenous mycobacterial MCE lipid transporter. Nature 620, 445–452. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 74. Wallner B (2023) AFsample: improving multimer prediction with AlphaFold using massive sampling. Bioinformatics 39, btad573. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 75. Abramson J, Adler J, Dunger J, Evans R, Green T, Pritzel A, Ronneberger O, Willmore L, Ballard AJ, Bambrick J et al. (2024) Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 630, 493–500. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 76. Shor B and Schneidman‐Duhovny D (2024) CombFold: predicting structures of large protein assemblies using a combinatorial assembly algorithm and AlphaFold2. Nat Methods 21, 477–487. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 77. Xia Y, Zhao K, Liu D, Zhou X and Zhang G (2023) Multi‐domain and complex protein structure prediction using inter‐domain interactions from deep learning. Commun Biol 6, 1221. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 78. Zheng W, Wuyun Q, Li Y, Zhang C, Freddolino PL and Zhang Y (2024) Improving deep learning protein monomer and complex structure prediction using DeepMSA2 with huge metagenomics data. Nat Methods 21, 279–289. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 79. Weissenow K, Heinzinger M, Steinegger M and Rost B (2022) Ultra‐fast protein structure prediction to capture effects of sequence variation in mutation movies. bioRxiv. doi: 10.1101/2022.11.14.516473 [DOI] [Google Scholar]
• 80. Lee JH, Yadollahpour P, Watkins A, Frey NC, Leaver‐Fay A, Ra S, Cho K, Gligorijevic V, Regev A and Bonneau R (2023) EquiFold: protein structure prediction with a novel coarse‐grained structure representation. bioRxiv. doi: 10.1101/2022.10.07.511322 [DOI] [Google Scholar]
• 81. Fang X, Wang F, Liu L, He J, Lin D, Xiang Y, Zhu K, Zhang X, Wu H, Li H et al. (2023) A method for multiple‐sequence‐alignment‐free protein structure prediction using a protein language model. Nat Mach Intell 5, 1087–1096. [Google Scholar]
• 82. Bryant P, Pozzati G, Zhu W, Shenoy A, Kundrotas P and Elofsson A (2022) Predicting the structure of large protein complexes using AlphaFold and Monte Carlo tree search. Nat Commun 13, 6028. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 83. Wu R, Ding F, Wang R, Shen R, Zhang X, Luo S, Su C, Wu Z, Xie Q, Berger B et al. (2022) High‐resolution de novo structure prediction from primary sequence. bioRxiv. doi: 10.1101/2022.07.21.500999 [DOI] [Google Scholar]
• 84. Baek M, DiMaio F, Anishchenko I, Dauparas J, Ovchinnikov S, Lee GR, Wang J, Cong Q, Kinch LN, Schaeffer RD et al. (2021) Accurate prediction of protein structures and interactions using a three‐track neural network. Science 373, 871–876. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 85. Krishna R, Wang J, Ahern W, Sturmfels P, Venkatesh P, Kalvet I, Lee GR, Morey‐Burrows FS, Anishchenko I, Humphreys IR et al. (2024) Generalized biomolecular modeling and design with RoseTTAFold All‐Atom. Science 384, eadl2528. [DOI] [PubMed] [Google Scholar]
• 86. Baek M, McHugh R, Anishchenko I, Jiang H, Baker D and DiMaio F (2024) Accurate prediction of protein–nucleic acid complexes using RoseTTAFoldNA. Nat Methods 21, 117–121. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 87. Bryant P, Kelkar A, Guljas A, Clementi C and Noé F (2024) Structure prediction of protein–ligand complexes from sequence information with Umol. Nat Commun 15, 4536. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 88. Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, Kaiser Ł and Polosukhin I (2017) Attention Is All You Need. 31st Conference on Neural Information Processing Systems (NIPS 2017), Long Beach, CA, USA.
• 89. Chowdhury R, Bouatta N, Biswas S, Floristean C, Kharkar A, Roy K, Rochereau C, Ahdritz G, Zhang J, Church GM et al. (2022) Single‐sequence protein structure prediction using a language model and deep learning. Nat Biotechnol 40, 1617–1623. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 90. Das R, Kretsch RC, Simpkin AJ, Mulvaney T, Pham P, Rangan R, Bu F, Keegan RM, Topf M, Rigden DJ et al. (2023) Assessment of three‐dimensional RNA structure prediction in CASP15. Proteins 91, 1747–1770. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 91. Robin X, Studer G, Durairaj J, Eberhardt J, Schwede T and Walters WP (2023) Assessment of protein–ligand complexes in CASP15. Proteins 91, 1811–1821. [DOI] [PubMed] [Google Scholar]
• 92. Cai H, Shen C, Jian T, Zhang X, Chen T, Han X, Yang Z, Dang W, Hsieh CY, Kang Y et al. (2023) CarsiDock: a deep learning paradigm for accurate protein–ligand docking and screening based on large‐scale pre‐training. Chem Sci 15, 1449–1471. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 93. Joosten RP, Salzemann J, Bloch V, Stockinger H, Berglund AC, Blanchet C, Bongcam‐Rudloff E, Combet C, Da Costa AL, Deleage G et al. (2009) PDB_REDO: automated re‐refinement of X‐ray structure models in the PDB. J Appl Cryst 42, 376–384. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 94. Fischer JD, Holliday GL and Thornton JM (2010) The CoFactor database: organic cofactors in enzyme catalysis. Bioinformatics 26, 2496–2497. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 95. Lee C, Su BH and Tseng YJ (2022) Comparative studies of AlphaFold, RoseTTAFold and Modeller: a case study involving the use of G‐protein‐coupled receptors. Brief Bioinform 23, bbac308. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 96. Gil Zuluaga FH, D'Arminio N, Bardozzo F, Tagliaferri R and Marabotti A (2023) An automated pipeline integrating AlphaFold 2 and MODELLER for protein structure prediction. Comput Struct Biotechnol J 21, 5620–5629. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 97. Zheng L, Shi S, Sun X, Lu M, Liao Y, Zhu S, Zhang H, Pan Z, Fang P, Zeng Z et al. (2024) MoDAFold: a strategy for predicting the structure of missense mutant protein based on AlphaFold2 and molecular dynamics. Brief Bioinform 25, bbae006. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 98. Lafita A, Bliven S, Kryshtafovych A, Bertoni M, Monastyrskyy B, Duarte JM, Schwede T and Capitani G (2018) Assessment of protein assembly prediction in CASP12. Proteins 86, 247–256. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 99. Ozden B, Kryshtafovych A and Karaca E (2021) Assessment of the CASP14 assembly predictions. Proteins 89, 1787–1799. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 100. Evans R, O'Neill M, Pritzel A, Antropova N, Senior A, Green T, Žídek A, Bates R, Blackwell S, Yim J et al. (2021) Protein complex prediction with AlphaFold‐Multimer. bioRxiv. doi: 10.1101/2021.10.04.463034v2 [DOI] [Google Scholar]
• 101. Elofsson A (2023) Progress at protein structure prediction, as seen in CASP15. Curr Opin Struct Biol 80, 102594. [DOI] [PubMed] [Google Scholar]
• 102. Liu J, Guo Z, Wu T, Roy RS, Chen C and Cheng J (2023) Improving AlphaFold2–based protein tertiary structure prediction with MULTICOM in CASP15. Commun Chem 6, 188. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 103. Pereira GP, Gouzien C, Souza PCT and Martin J (2024) AlphaFold‐Multimer struggles in predicting PROTAC‐mediated protein–protein interfaces. bioRxiv. doi: 10.1101/2024.03.19.585735 [DOI] [Google Scholar]
• 104. Liu Y, Yang J, Wang T, Luo M, Chen Y, Chen C, Ronai Z, Zhou Y, Ruppin E and Han L (2023) Expanding PROTACtable genome universe of E3 ligases. Nat Commun 14, 6509. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 105. Zaidman D, Prilusky J and London N (2020) PRosettaC: Rosetta based modeling of PROTAC mediated ternary complexes. J Chem Inf Model 60, 4894–4903. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 106. Chiu TP, Rao S and Rohs R (2023) Physicochemical models of protein–DNA binding with standard and modified base pairs. Proc Natl Acad Sci USA 120, e2205796120. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 107. Zhao L, Koseki SRT, Silverstein RA, Amrani N, Peng C, Kramme C, Savic N, Pacesa M, Rodríguez TC, Stan T et al. (2023) PAM‐flexible genome editing with an engineered chimeric Cas9. Nat Commun 14, 6175. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 108. Li J, Chiu TP and Rohs R (2024) Predicting DNA structure using a deep learning method. Nat Commun 15, 1243. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 109. Shulgina Y, Trinidad MI, Langeberg CJ, Nisonoff H, Chithrananda S, Skopintsev P, Nissley AJ, Patel J, Boger RS, Shi H et al. (2024) RNA language models predict mutations that improve RNA function. bioRxiv. doi: 10.1101/2024.04.05.588317 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 110. Gong T and Bu D (2024) Language models enable zero‐shot prediction of RNA secondary structure including pseudoknots. bioRxiv. doi: 10.1101/2024.01.27.577533 [DOI] [Google Scholar]
• 111. Nguyen E, Poli M, Durrant MG, Thomas AW, Kang B, Sullivan J, Ng MY, Lewis A, Patel A, Lou A et al. (2024) Sequence modeling and design from molecular to genome scale with Evo. bioRxiv. doi: 10.1101/2024.02.27.582234 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 112. Jisna VA and Jayaraj PB (2021) Protein structure prediction: conventional and deep learning perspectives. Protein J 40, 522–544. [DOI] [PubMed] [Google Scholar]
• 113. Lane TJ (2023) Protein structure prediction has reached the single‐structure frontier. Nat Methods 20, 170–173. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 114. Sala D, Engelberger F, Mchaourab HS and Meiler J (2023) Modeling conformational states of proteins with AlphaFold. Curr Opin Struct Biol 81, 102645. [DOI] [PubMed] [Google Scholar]
• 115. Nugent T, Cozzetto D and Jones DT (2014) Evaluation of predictions in the CASP10 model refinement category. Proteins 82, 98–111. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 116. Sala D, Hildebrand PW and Meiler J (2023) Biasing AlphaFold2 to predict GPCRs and kinases with user‐defined functional or structural properties. Front Mol Biosci 10, 1121962. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 117. Del Alamo D, Sala D, Mchaourab HS and Meiler J (2022) Sampling alternative conformational states of transporters and receptors with AlphaFold2. Elife 11, e75751. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 118. Heo L and Feig M (2022) Multi‐state modeling of G‐protein coupled receptors at experimental accuracy. Proteins 90, 1873–1885. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 119. Stein RA and Mchaourab HS (2022) SPEACH_AF: sampling protein ensembles and conformational heterogeneity with Alphafold2. PLoS Comput Biol 18, e1010483. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 120. Wayment‐Steele HK, Ojoawo A, Otten R, Apitz JM, Pitsawong W, Hömberger M, Ovchinnikov S, Colwell L and Kern D (2024) Predicting multiple conformations via sequence clustering and AlphaFold2. Nature 625, 832–839. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 121. Jing B, Berger B and Jaakkola T (2024) AlphaFold meets flow matching for generating protein ensembles. arXiv: 2402.04845.
• 122. Yim J, Campbell A, Mathieu E, Foong AYK, Gastegger M, Jiménez‐Luna J, Lewis S, Satorras VG, Veeling BS, Noé F et al. (2024) Improved motif‐scaffolding with SE(3) flow matching. arXiv: 2401.04082v1.
• 123. Janson G, Valdes‐Garcia G, Heo L and Feig M (2023) Direct generation of protein conformational ensembles via machine learning. Nat Commun 14, 774. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 124. Masrati G, Landau M, Ben‐Tal N, Lupas A, Kosloff M and Kosinski J (2021) Integrative structural biology in the era of accurate structure prediction. J Mol Biol 433, 167127. [DOI] [PubMed] [Google Scholar]
• 125. Regan L, Caballero D, Hinrichsen MR, Virrueta A, Williams DM and O'Hern CS (2015) Protein design: past, present, and future. Biopolymers 104, 334–350. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 126. Qureshi R, Irfan M, Gondal TM, Khan S, Wu J, Hadi MU, Heymach J, Le X, Yan H and Alam T (2023) AI in drug discovery and its clinical relevance. Heliyon 9, e17575. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 127. Harrer S, Menard J, Rivers M, Green DVS, Karpiak J, Jeliazkov R, Shapovalov MV, del Alamo D and Sternke MC (2024) Artificial intelligence drives the digital transformation of pharma. In Artificial Intelligence in Clinical Practice (Krittanawong C, ed.), pp. 345–372. Academic Press, Cambridge, MA. [Google Scholar]
• 128. Almubarak HF, Tan W, Hoffmann AD, Wei J, El‐Shennawy L, Squires JR, Sun Y, Dashzeveg NK, Simonton B, Jia Y et al. (2024) Physics‐driven structural docking and protein language models accelerate antibody screening and design for broad‐spectrum antiviral therapy. bioRxiv. doi: 10.1101/2024.03.01.582176 [DOI] [Google Scholar]
• 129. No authors listed (2024) Spotlight on protein structure design. Nat Biotechnol 42, 157. [DOI] [PubMed] [Google Scholar]
• 130. Ruffolo JA, Chu LS, Mahajan SP and Gray JJ (2023) Fast, accurate antibody structure prediction from deep learning on massive set of natural antibodies. Nat Commun 14, 2389. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 131. Ingraham JB, Baranov M, Costello Z, Barber KW, Wang W, Ismail A, Frappier V, Lord DM, Ng‐Thow‐Hing C, Van Vlack ER et al. (2023) Illuminating protein space with a programmable generative model. Nature 623, 1070–1078. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 132. Ruffolo JA, Sulam J and Gray JJ (2021) Antibody structure prediction using interpretable deep learning. Patterns 3, 100406. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 133. Ruffolo JA, Guerra C, Mahajan SP, Sulam J and Gray JJ (2020) Geometric potentials from deep learning improve prediction of CDR H3 loop structures. Bioinformatics 36, i268–i275. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 134. Jing B, Erives E, Pao‐Huang P, Corso G, Berger B and Jaakkola T (2023) EigenFold: generative protein structure prediction with diffusion models. arXiv: 2304.02198.
• 135. Hayes T, Rao R, Akin H, Sofroniew NJ, Oktay D, Lin Z, Verkuil R, Tran VQ, Deaton J, Wiggert M et al. (2024) Simulating 500 million years of evolution with a language model. bioRxiv. doi: 10.1101/2024.07.01.600583 [DOI] [PubMed] [Google Scholar]
• 136. Alamdari S, Thakkar N, van den Berg R, Lu AX, Fusi N, Amini AP and Yang KK (2023) Protein generation with evolutionary diffusion: sequence is all you need. bioRxiv. doi: 10.1101/2023.09.11.556673 [DOI] [Google Scholar]
• 137. Wu KE, Yang KK, Berg R, van den Zou JY, Lu AX and Amini AP (2022) Protein structure generation via folding diffusion. arXiv: 2209.15611. [DOI] [PMC free article] [PubMed]
• 138. Zhang C, Leach A, Makkink T, Arbesú M, Kadri I, Luo D, Mizrahi L, Krichen S, Lang M, Tovchigrechko A et al. (2023) FrameDiPT: SE(3) diffusion model for protein structure inpainting. bioRxiv. doi: 10.1101/2023.11.21.568057 [DOI] [Google Scholar]
• 139. Lin Y and AlQuraishi M (2023) Generating novel, designable, and diverse protein structures by equivariantly diffusing oriented residue clouds. arXiv: 2301.12485.
• 140. Chen S, Lin T, Basu R, Ritchey J, Wang S, Luo Y, Li X, Pei D, Kara LB and Cheng X (2024) Design of target‐specific peptide inhibitors using generative deep learning and molecular dynamics simulations. Nat Commun 15, 1611. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 141. Xie X, Valiente PA, Kim J and Kim PM (2024) HelixDiff, a score‐based diffusion model for generating all‐atom α‐helical structures. ACS Cent Sci 10, 1001–1011. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 142. Xie X, Valiente PA and Kim PM (2023) HelixGAN a deep‐learning methodology for conditional de novo design of α‐helix structures. Bioinformatics 39, btad036. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 143. Gainza P, Wehrle S, Van Hall‐Beauvais A, Marchand A, Scheck A, Harteveld Z, Buckley S, Ni D, Tan S, Sverrisson F et al. (2023) De novo design of protein interactions with learned surface fingerprints. Nature 617, 176–184. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 144. Watson JL, Juergens D, Bennett NR, Trippe BL, Yim J, Eisenach HE, Ahern W, Borst AJ, Ragotte RJ, Milles LF et al. (2023) De novo design of protein structure and function with RFdiffusion. Nature 620, 1089–1100. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 145. Kortemme T (2024) De novo protein design‐from new structures to programmable functions. Cell 187, 526–544. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 146. Yim J, Stärk H, Corso G, Jing B, Barzilay R and Jaakkola TS (2024) Diffusion models in protein structure and docking. Wiley Interdiscip Rev Comput Mol Sci 14, e1711. [Google Scholar]
• 147. Wang XF, Tang JY, Liang H, Sun J, Dorje S, Peng B, Ji XW, Li Z, Zhang XE and Wang DB (2024) ProT–Diff: a modularized and efficient approach to de novo generation of antimicrobial peptide sequences through integration of protein language model and diffusion model. bioRxiv. doi: 10.1101/2024.02.22.581480 [DOI] [PMC free article] [PubMed] [Google Scholar]
• 148. Ho J, Jain A and Abbeel P (2020) Denoising diffusion probabilistic models. Adv Neural Inf Process Syst 33, 6840–6851. [Google Scholar]
• 149. Dhariwal P and Nichol A (2021) Diffusion models beat GANs on image synthesis. Adv Neural Inf Process Syst 34, 8780–8794. [Google Scholar]
• 150. Guo Z, Liu J, Wang Y, Chen M, Wang D, Xu D and Cheng J (2024) Diffusion models in bioinformatics and computational biology. Nat Rev Bioeng 2, 136–154. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 151. Lin Y, Lee M, Zhang Z and AlQuraishi M (2024) Out of many, one: designing and scaffolding proteins at the scale of the structural universe with Genie 2. arXiv: 2405.15489.
• 152. Dauparas J, Anishchenko I, Bennett N, Bai H, Ragotte RJ, Milles LF, Wicky BIM, Courbet A, de Haas RJ, Bethel N et al. (2022) Robust deep learning‐based protein sequence design using ProteinMPNN. Science 378, 49–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 153. Dauparas J, Lee GR, Pecoraro R, An L, Anishchenko I, Glasscock C and Baker D (2023) Atomic context‐conditioned protein sequence design using LigandMPNN. bioRxiv. doi: 10.1101/2023.12.22.573103 [DOI] [Google Scholar]
• 154. Chen H, Fan X, Zhu S, Pei Y, Zhang X, Zhang X, Liu L, Qian F and Tian B (2024) Accurate prediction of CDR‐H3 loop structures of antibodies with deep learning. Elife 12, RP91512. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 155. Pantazes RJ and Maranas CD (2010) OptCDR: a general computational method for the design of antibody complementarity determining regions for targeted epitope binding. Protein Eng Des Sel 23, 849–858. [DOI] [PubMed] [Google Scholar]
• 156. Sircar A, Kim ET and Gray JJ (2009) RosettaAntibody: antibody variable region homology modeling server. Nucleic Acids Res 37, W474–W479. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 157. Lipsh‐Sokolik R, Listov D and Fleishman SJ (2021) The AbDesign computational pipeline for modular backbone assembly and design of binders and enzymes. Protein Sci 30, 151–159. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 158. Abanades B, Georges G, Bujotzek A and Deane CM (2022) ABlooper: fast accurate antibody CDR loop structure prediction with accuracy estimation. Bioinformatics 38, 1877–1880. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 159. Yin R and Pierce BG (2024) Evaluation of AlphaFold antibody‐antigen modeling with implications for improving predictive accuracy. Protein Sci 33, e4865. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 160. Polonsky K, Pupko T and Freund NT (2023) Evaluation of the ability of AlphaFold to predict the three‐dimensional structures of antibodies and epitopes. J Immunol 211, 1578–1588. [DOI] [PubMed] [Google Scholar]
• 161. Xu Z, Davila A, Wilamowski J, Teraguchi S and Standley DM (2022) Improved antibody‐specific epitope prediction using AlphaFold and AbAdapt. Chembiochem 23, e202200303. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 162. Schoeder CT, Schmitz S, Adolf‐Bryfogle J, Sevy AM, Finn JA, Sauer MF, Bozhanova NG, Mueller BK, Sangha AK, Bonet J et al. (2021) Modeling immunity with Rosetta: methods for antibody and antigen design. Biochemistry 60, 825–846. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 163. Khan A, Cowen‐Rivers AI, Grosnit A, Deik DG, Robert PA, Greiff V, Smorodina E, Rawat P, Akbar R, Dreczkowski K et al. (2023) Toward real‐world automated antibody design with combinatorial Bayesian optimization. Cell Rep Methods 3, 100374. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 164. Eguchi RR, Choe CA, Parekh U, Khalek IS, Ward MD, Vithani N, Bowman GR, Jardine JG and Huang PS (2022) Deep generative design of epitope‐specific binding proteins by latent conformation optimization. bioRxiv. doi: 10.1101/2022.12.22.521698 [DOI] [Google Scholar]
• 165. Chang L, Mondal A and Perez A (2022) Towards rational computational peptide design. Front Bioinform 2, 1046493. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 166. Chen T, Pertsemlidis S, Watson R, Kavirayuni VS, Hsu A, Vure P, Pulugurta R, Vincoff S, Hong L, Wang T et al. (2023) PepMLM: target sequence‐conditioned generation of peptide binders via masked language modeling. arXiv: 2310.03842v2.
• 167. Li G, Iyer B, Prasath VBS, Ni Y and Salomonis N (2021) DeepImmuno: deep learning‐empowered prediction and generation of immunogenic peptides for T‐cell immunity. Brief Bioinform 22, bbab160. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 168. Van Oort CM, Ferrell JB, Remington JM, Wshah S and Li J (2021) AMPGAN v2: machine learning‐guided design of antimicrobial peptides. J Chem Inf Model 61, 2198–2207. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 169. Zervou MA, Doutsi E, Pantazis Y and Tsakalides P (2024) De novo antimicrobial peptide design with feedback generative adversarial networks. Int J Mol Sci 25, 5506. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 170. Surana S, Arora P, Singh D, Sahasrabuddhe D and Valadi J (2023) PandoraGAN: generating antiviral peptides using generative adversarial network. SN Comput Sci 4, 607. doi: 10.1007/s42979-023-02203-3 [DOI] [Google Scholar]
• 171. Chen T, Vure P, Pulugurta R and Chatterjee P (2024) AMP‐diffusion: integrating latent diffusion with protein language models for antimicrobial peptide generation. bioRxiv. doi: 10.1101/2024.03.03.583201 [DOI] [Google Scholar]
• 172. Bolchini D, Finkelstein A, Perrone V and Nagl S (2009) Better bioinformatics through usability analysis. Bioinformatics 25, 406–412. [DOI] [PubMed] [Google Scholar]
• 173. Goldman M and Gehlenborg N (2021) Making tools that people will use: user‐centered design in computational biology research. Pac Symp Biocomput 26, 346–350. [PubMed] [Google Scholar]
• 174. Pavelin K, Cham JA, de Matos P, Brooksbank C, Cameron G and Steinbeck C (2012) Bioinformatics meets user‐centred design: a perspective. PLoS Comput Biol 8, e1002554. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 175. Javahery H, Seffah A and Radhakrishnan T (2004) Beyond power: making bioinformatics tools user‐centered. Commun ACM 47, 59–62. [Google Scholar]
• 176. Ison J, Ienasescu H, Chmura P, Rydza E, Ménager H, Kalaš M, Schwämmle V, Grüning B, Beard N, Lopez R et al. (2019) The bio.tools registry of software tools and data resources for the life sciences. Genome Biol 20, 164. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 177. Kelley LA, Mezulis S, Yates CM, Wass MN and Sternberg MJ (2015) The Phyre2 web portal for protein modeling, prediction and analysis. Nat Protoc 10, 845–858. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 178. Schwede T, Kopp J, Guex N and Peitsch MC (2003) SWISS‐MODEL: an automated protein homology‐modeling server. Nucleic Acids Res 31, 3381–3385. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 179. Janson G and Paiardini A (2021) PyMod 3: a complete suite for structural bioinformatics in PyMOL. Bioinformatics 37, 1471–1472. [DOI] [PubMed] [Google Scholar]
• 180. Zhou Z, Hu M, Salcedo M, Gravel N, Yeung W, Venkat A, Guo D, Zhang J, Kannan N and Li S (2023) XAI meets biology: a comprehensive review of explainable AI in bioinformatics applications. arXiv: 2312.06082.
• 181. Fiser A (2010) Template‐based protein structure modeling. Methods Mol Biol 673, 73–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 182. Fraenkel AS (1993) Complexity of protein folding. Bull Math Biol 55, 1199–1210. [DOI] [PubMed] [Google Scholar]
• 183. Horsfall D, Cool J, Hettrick S, Pisco AO, Hong NC and Haniffa M (2023) Research software engineering accelerates the translation of biomedical research for health. Nat Med 29, 1313–1316. [DOI] [PubMed] [Google Scholar]
• 184. Pierce NA and Winfree E (2002) Protein design is NP‐hard. Protein Eng 15, 779–782. [DOI] [PubMed] [Google Scholar]
• 185. Zhong B, Su X, Wen M, Zuo S, Hong L and Lin L (2021) ParaFold: paralleling AlphaFold for large‐scale predictions. arXiv: 2111.06340. doi: 10.48550/arXiv.2111.06340 [DOI]
• 186. MacCarthy EA, Zhang C, Zhang Y and Kc DB (2022) GPU‐I‐TASSER: a GPU accelerated I‐TASSER protein structure prediction tool. Bioinformatics 38, 1754–1755. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 187. Bertoline LMF, Lima AN, Krieger JE and Teixeira SK (2023) Before and after AlphaFold2: an overview of protein structure prediction. Front Bioinform 3, 1120370. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 188. Xu T, Xu Q and Li J (2023) Toward the appropriate interpretation of Alphafold2. Front Artif Intell 6, 1149748. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 189. Pakhrin SC, Shrestha B, Adhikari B and Kc DB (2021) Deep learning‐based advances in protein structure prediction. Int J Mol Sci 22, 5553. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 190. Aithani L, Alcaide E, Bartunov S, Cooper CDO, Doré AS, Lane TJ, Maclean F, Rucktooa P, Shaw RA and Skerratt SE (2023) Advancing structural biology through breakthroughs in AI. Curr Opin Struct Biol 80, 102601. [DOI] [PubMed] [Google Scholar]
• 191. Borkakoti N and Thornton JM (2023) AlphaFold2 protein structure prediction: implications for drug discovery. Curr Opin Struct Biol 78, 102526. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 192. Khatami MH, Mendes UC, Wiebe N and Kim PM (2023) Gate‐based quantum computing for protein design. PLoS Comput Biol 19, e1011033. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 193. Mulligan VK, Melo H, Merritt HI, Slocum S, Weitzner BD, Watkins AM, Renfrew PD, Pelissier C, Arora PS and Bonneau R (2020) Designing peptides on a quantum computer. bioRxiv. doi: 10.1101/752485 [DOI] [Google Scholar]