AlphaFold at CASP13

The importance and future of biochemical engineering

Today's Biochemical Engineer may contribute to advances in a wide range of technical areas. The recent Biochemical and Molecular Engineering XXI conference focused on "The Next Generation of Biochemical and Molecular Engineering: The role of emerging technologies in tomorrow's products and processes". On the basis of topical discussions at this conference, this perspective synthesizes one vision on where investment in research areas is needed for biotechnology to continue contributing to some of the world's grand challenges.

Residue Cluster Classes: A Unified Protein Representation for Efficient Structural and Functional Classification

Proteins are characterized by their structures and functions, and these two fundamental aspects of proteins are assumed to be related. To model such a relationship, a single representation to model both protein structure and function would be convenient, yet so far, the most effective models for protein structure or function classification do not rely on the same protein representation. Here we provide a computationally efficient implementation for large datasets to calculate residue cluster classes (RCCs) from protein three-dimensional structures and show that such representations enable a random forest algorithm to effectively learn the structural and functional classifications of proteins, according to the CATH and Gene Ontology criteria, respectively. RCCs are derived from residue contact maps built from different distance criteria, and we show that 7 or 8 Å with or without amino acid side-chain atoms rendered the best classification models. The potential use of a unified representation of proteins is discussed and possible future areas for improvement and exploration are presented.

UDSMProt: universal deep sequence models for protein classification

MOTIVATION

Inferring the properties of a protein from its amino acid sequence is one of the key problems in bioinformatics. Most state-of-the-art approaches for protein classification are tailored to single classification tasks and rely on handcrafted features, such as position-specific-scoring matrices from expensive database searches. We argue that this level of performance can be reached or even be surpassed by learning a task-agnostic representation once, using self-supervised language modeling, and transferring it to specific tasks by a simple fine-tuning step.

RESULTS

We put forward a universal deep sequence model that is pre-trained on unlabeled protein sequences from Swiss-Prot and fine-tuned on protein classification tasks. We apply it to three prototypical tasks, namely enzyme class prediction, gene ontology prediction and remote homology and fold detection. The proposed method performs on par with state-of-the-art algorithms that were tailored to these specific tasks or, for two out of three tasks, even outperforms them. These results stress the possibility of inferring protein properties from the sequence alone and, on more general grounds, the prospects of modern natural language processing methods in omics. Moreover, we illustrate the prospects for explainable machine learning methods in this field by selected case studies.

AVAILABILITY AND IMPLEMENTATION

Source code is available under https://github.com/nstrodt/UDSMProt.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Protein Interface Complementarity and Gene Duplication Improve Link Prediction of Protein-Protein Interaction Network

Protein-protein interactions are the foundations of cellular life activities. At present, the already known protein-protein interactions only account for a small part of the total. With the development of experimental and computing technology, more and more PPI data are mined, PPI networks are more and more dense. It is possible to predict protein-protein interaction from the perspective of network structure. Although there are many high-throughput experimental methods to detect protein-protein interactions, the cost of experiments is high, time-consuming, and there is a certain error rate meanwhile. Network-based approaches can provide candidates of protein pairs for high-throughput experiments and improve the accuracy rate. This paper presents a new link prediction approach "Sim" for PPI networks from the perspectives of proteins' complementary interfaces and gene duplication. By integrating our approach "Sim" with the state-of-art network-based approach "L3," the prediction accuracy and robustness are improved.

Big data and artificial intelligence discover novel drugs targeting proteins without 3D structure and overcome the undruggable targets

The discovery of targeted drugs heavily relies on three-dimensional (3D) structures of target proteins. When the 3D structure of a protein target is unknown, it is very difficult to design its corresponding targeted drugs. Although the 3D structures of some proteins (the so-called undruggable targets) are known, their targeted drugs are still absent. As increasing crystal/cryogenic electron microscopy structures are deposited in Protein Data Bank, it is much more possible to discover the targeted drugs. Moreover, it is also highly probable to turn previous undruggable targets into druggable ones when we identify their hidden allosteric sites. In this review, we focus on the currently available advanced methods for the discovery of novel compounds targeting proteins without 3D structure and how to turn undruggable targets into druggable ones.

Neighborhood Preference of Amino Acids in Protein Structures and its Applications in Protein Structure Assessment

Amino acids form protein 3D structures in unique manners such that the folded structure is stable and functional under physiological conditions. Non-specific and non-covalent interactions between amino acids exhibit neighborhood preferences. Based on structural information from the protein data bank, a statistical energy function was derived to quantify amino acid neighborhood preferences. The neighborhood of one amino acid is defined by its contacting residues, and the energy function is determined by the neighboring residue types and relative positions. The neighborhood preference of amino acids was exploited to facilitate structural quality assessment, which was implemented in the neighborhood preference program NEPRE. The source codes are available via https://github.com/LiuLab-CSRC/NePre.

The Application of Deep Learning in Cancer Prognosis Prediction

Deep learning has been applied to many areas in health care, including imaging diagnosis, digital pathology, prediction of hospital admission, drug design, classification of cancer and stromal cells, doctor assistance, etc. Cancer prognosis is to estimate the fate of cancer, probabilities of cancer recurrence and progression, and to provide survival estimation to the patients. The accuracy of cancer prognosis prediction will greatly benefit clinical management of cancer patients. The improvement of biomedical translational research and the application of advanced statistical analysis and machine learning methods are the driving forces to improve cancer prognosis prediction. Recent years, there is a significant increase of computational power and rapid advancement in the technology of artificial intelligence, particularly in deep learning. In addition, the cost reduction in large scale next-generation sequencing, and the availability of such data through open source databases (e.g., TCGA and GEO databases) offer us opportunities to possibly build more powerful and accurate models to predict cancer prognosis more accurately. In this review, we reviewed the most recent published works that used deep learning to build models for cancer prognosis prediction. Deep learning has been suggested to be a more generic model, requires less data engineering, and achieves more accurate prediction when working with large amounts of data. The application of deep learning in cancer prognosis has been shown to be equivalent or better than current approaches, such as Cox-PH. With the burst of multi-omics data, including genomics data, transcriptomics data and clinical information in cancer studies, we believe that deep learning would potentially improve cancer prognosis.

A review of mathematical representations of biomolecular data

Recently, machine learning (ML) has established itself in various worldwide benchmarking competitions in computational biology, including Critical Assessment of Structure Prediction (CASP) and Drug Design Data Resource (D3R) Grand Challenges. However, the intricate structural complexity and high ML dimensionality of biomolecular datasets obstruct the efficient application of ML algorithms in the field. In addition to data and algorithm, an efficient ML machinery for biomolecular predictions must include structural representation as an indispensable component. Mathematical representations that simplify the biomolecular structural complexity and reduce ML dimensionality have emerged as a prime winner in D3R Grand Challenges. This review is devoted to the recent advances in developing low-dimensional and scalable mathematical representations of biomolecules in our laboratory. We discuss three classes of mathematical approaches, including algebraic topology, differential geometry, and graph theory. We elucidate how the physical and biological challenges have guided the evolution and development of these mathematical apparatuses for massive and diverse biomolecular data. We focus the performance analysis on protein-ligand binding predictions in this review although these methods have had tremendous success in many other applications, such as protein classification, virtual screening, and the predictions of solubility, solvation free energies, toxicity, partition coefficients, protein folding stability changes upon mutation, etc.

Modeling to Understand Plant Protein Structure-Function Relationships-Implications for Seed Storage Proteins

Proteins are among the most important molecules on Earth. Their structure and aggregation behavior are key to their functionality in living organisms and in protein-rich products. Innovations, such as increased computer size and power, together with novel simulation tools have improved our understanding of protein structure-function relationships. This review focuses on various proteins present in plants and modeling tools that can be applied to better understand protein structures and their relationship to functionality, with particular emphasis on plant storage proteins. Modeling of plant proteins is increasing, but less than 9% of deposits in the Research Collaboratory for Structural Bioinformatics Protein Data Bank come from plant proteins. Although, similar tools are applied as in other proteins, modeling of plant proteins is lagging behind and innovative methods are rarely used. Molecular dynamics and molecular docking are commonly used to evaluate differences in forms or mutants, and the impact on functionality. Modeling tools have also been used to describe the photosynthetic machinery and its electron transfer reactions. Storage proteins, especially in large and intrinsically disordered prolamins and glutelins, have been significantly less well-described using modeling. These proteins aggregate during processing and form large polymers that correlate with functionality. The resulting structure-function relationships are important for processed storage proteins, so modeling and simulation studies, using up-to-date models, algorithms, and computer tools are essential for obtaining a better understanding of these relationships.

Improved Consensus-Fragment Selection in Template-Assisted Prediction of Protein Structures with the UNRES Force Field in CASP13

The method for protein-structure prediction, which combines the physics-based coarse-grained UNRES force field with knowledge-based modeling, has been developed further and tested in the 13th Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction (CASP13). The method implements restraints from the consensus fragments common to server models. In this work, the server models to derive fragments have been chosen on the basis of quality assessment; a fully automatic fragment-selection procedure has been introduced, and Dynamic Fragment Assembly pseudopotentials have been fully implemented. The Global Distance Test Score (GDT_TS), averaged over our “Model 1” predictions, increased by over 10 units with respect to CASP12 for the free-modeling category to reach 40.82. Our “Model 1” predictions ranked 20 and 14 for all and free-modeling targets, respectively (upper 20.2% and 14.3% of all models submitted to CASP13 in these categories, respectively), compared to 27 (upper 21.1%) and 24 (upper 18.9%) in CASP12, respectively. For oligomeric targets, the Interface Patch Similarity (IPS) and Interface Contact Similarity (ICS) averaged over our best oligomer models increased from 0.28 to 0.36 and from 12.4 to 17.8, respectively, from CASP12 to CASP13, and top-ranking models of 2 targets (H0968 and T0997o) were obtained (none in CASP12). The improvement of our method in CASP13 over CASP12 was ascribed to the combined effect of the overall enhancement of server-model quality, our success in selecting server models and fragments to derive restraints, and improvements of the restraint and potential-energy functions.

A Benchmark for Homomeric Enzyme Active Site Structure Prediction Highlights the Importance of Accurate Modeling of Protein Symmetry

Accurate prediction and modeling of an enzyme's active site are critical for engineering efforts as well as providing insight into an enzyme's naturally occurring function. Previous efforts demonstrated that the integration of constraints enforcing strict geometric orientations between catalytic residues significantly improved the modeling accuracy for the active sites of monomeric enzymes. In this study, a similar approach was explored to evaluate the effect on the active sites of homomeric enzymes. A benchmark of 17 homomeric enzymes with known structures and a bound ligand relevant to the established chemistry were identified from the protein data bank. The enzymes identified span multiple classes as well as symmetries. Unlike what was observed for the monomeric enzymes, upon the application of catalytic geometric constraints, there was no significant improvement observed in modeling accuracy for either the active site of the protein structure or the accuracy of the subsequently docked ligand. Upon further analysis, it is apparent that the symmetric interface being modeled is inaccurate and prevented the active sites from being modeled at atomic-level accuracy. This is consistent with the challenge others have identified in being able to predict de novo protein symmetry. To further improve the accuracy of active site modeling for homomeric proteins, new methodologies to accurately model the symmetric interfaces of these complexes are needed.

Structure prediction of biological assemblies using GALAXY in CAPRI rounds 38-45

We participated in CARPI rounds 38-45 both as a server predictor and a human predictor. These CAPRI rounds provided excellent opportunities for testing prediction methods for three classes of protein interactions, that is, protein-protein, protein-peptide, and protein-oligosaccharide interactions. Both template-based methods (GalaxyTBM for monomer protein, GalaxyHomomer for homo-oligomer protein, GalaxyPepDock for protein-peptide complex) and ab initio docking methods (GalaxyTongDock and GalaxyPPDock for protein oligomer, GalaxyPepDock-ab-initio for protein-peptide complex, GalaxyDock2 and Galaxy7TM for protein-oligosaccharide complex) have been tested. Template-based methods depend heavily on the availability of proper templates and template-target similarity, and template-target difference is responsible for inaccuracy of template-based models. Inaccurate template-based models could be improved by our structure refinement and loop modeling methods based on physics-based energy optimization (GalaxyRefineComplex and GalaxyLoop) for several CAPRI targets. Current ab initio docking methods require accurate protein structures as input. Small conformational changes from input structure could be accounted for by our docking methods, producing one of the best models for several CAPRI targets. However, predicting large conformational changes involving protein backbone is still challenging, and full exploration of physics-based methods for such problems is still to come.

Assessing the accuracy of contact predictions in CASP13

The accuracy of sequence-based tertiary contact predictions was assessed in a blind prediction experiment at the CASP13 meeting. After 4 years of significant improvements in prediction accuracy, another dramatic advance has taken place since CASP12 was held 2 years ago. The precision of predicting the top L/5 contacts in the free modeling category, where L is the corresponding length of the protein in residues, has exceeded 70%. As a comparison, the best-performing group at CASP12 with a 47% precision would have finished below the top 1/3 of the CASP13 groups. Extensively trained deep neural network approaches dominate the top performing algorithms, which appear to efficiently integrate information on coevolving residues and interacting fragments or possibly utilize memories of sequence similarities and sometimes can deliver accurate results even in the absence of virtually any target specific evolutionary information. If the current performance is evaluated by F-score on L contacts, it stands around 24% right now, which, despite the tremendous impact and advance in improving its utility for structure modeling, also suggests that there is much room left for further improvement.

High-accuracy protein structures by combining machine-learning with physics-based refinement

Protein structure prediction has long been available as an alternative to experimental structure determination, especially via homology modeling based on templates from related sequences. Recently, models based on distance restraints from coevolutionary analysis via machine learning to have significantly expanded the ability to predict structures for sequences without templates. One such method, AlphaFold, also performs well on sequences where templates are available but without using such information directly. Here we show that combining machine-learning based models from AlphaFold with state-of-the-art physics-based refinement via molecular dynamics simulations further improves predictions to outperform any other prediction method tested during the latest round of CASP. The resulting models have highly accurate global and local structures, including high accuracy at functionally important interface residues, and they are highly suitable as initial models for crystal structure determination via molecular replacement.

A further leap of improvement in tertiary structure prediction in CASP13 prompts new routes for future assessments

We present our assessment of tertiary structure predictions for hard targets in Critical Assessment of Structure Prediction round 13 (CASP13). The analysis includes (a) assignment and discussion of best models through scores-aided visual inspection of models for each evaluation unit (EU); (b) ranking of predictors resulting from this evaluation and from global scores; and (c) evaluation of progress, state of the art, and current limitations of protein structure prediction. We witness a sizable improvement in tertiary structure prediction building on the progress observed from CASP11 to CASP12, with (a) top models reaching backbone RMSD <3 å for several EUs of size <150 residues, contributed by many groups; (b) at least one model that roughly captures global topology for all EUs, probably unprecedented in this track of CASP; and (c) even quite good models for full, unsplit targets. Better structure predictions are brought about mainly by improved residue-residue contact predictions, and since this CASP also by distance predictions, achieved through state-of-the-art machine learning methods which also progressed to work with slightly shallower alignments compared to CASP12. As we reach a new realm of tertiary structure prediction quality, new directions are proposed and explored for future CASPs: (a) dropping splitting into EUs, (b) rethinking difficulty metrics probably in terms of contact and distance predictions, (c) assessing also side chains for models of high backbone accuracy, and (d) assessing residue-wise and possibly residue-residue quality estimates.

Protein structure prediction beyond AlphaFold

DeepMind's AlphaFold recently demonstrated the potential of deep learning for protein structure prediction. DeepFragLib, a new protein-specific fragment library built using deep neural networks, may have advanced the field to the next stage.