ASTRO-FOLD 2.0: an Enhanced Framework for Protein Structure Prediction

A Guide to In Silico Drug Design

The drug discovery process is a rocky path that is full of challenges, with the result that very few candidates progress from hit compound to a commercially available product, often due to factors, such as poor binding affinity, off-target effects, or physicochemical properties, such as solubility or stability. This process is further complicated by high research and development costs and time requirements. It is thus important to optimise every step of the process in order to maximise the chances of success. As a result of the recent advancements in computer power and technology, computer-aided drug design (CADD) has become an integral part of modern drug discovery to guide and accelerate the process. In this review, we present an overview of the important CADD methods and applications, such as in silico structure prediction, refinement, modelling and target validation, that are commonly used in this area.

COMTOP: Protein Residue-Residue Contact Prediction through Mixed Integer Linear Optimization

Protein contact prediction helps reconstruct the tertiary structure that greatly determines a protein’s function; therefore, contact prediction from the sequence is an important problem. Recently there has been exciting progress on this problem, but many of the existing methods are still low quality of prediction accuracy. In this paper, we present a new mixed integer linear programming (MILP)-based consensus method: a Consensus scheme based On a Mixed integer linear opTimization method for prOtein contact Prediction (COMTOP). The MILP-based consensus method combines the strengths of seven selected protein contact prediction methods, including CCMpred, EVfold, DeepCov, NNcon, PconsC4, plmDCA, and PSICOV, by optimizing the number of correctly predicted contacts and achieving a better prediction accuracy. The proposed hybrid protein residue–residue contact prediction scheme was tested in four independent test sets. For 239 highly non-redundant proteins, the method showed a prediction accuracy of 59.68%, 70.79%, 78.86%, 89.04%, 94.51%, and 97.35% for top-5L, top-3L, top-2L, top-L, top-L/2, and top-L/5 contacts, respectively. When tested on the CASP13 and CASP14 test sets, the proposed method obtained accuracies of 75.91% and 77.49% for top-L/5 predictions, respectively. COMTOP was further tested on 57 non-redundant α-helical transmembrane proteins and achieved prediction accuracies of 64.34% and 73.91% for top-L/2 and top-L/5 predictions, respectively. For all test datasets, the improvement of COMTOP in accuracy over the seven individual methods increased with the increasing number of predicted contacts. For example, COMTOP performed much better for large number of contact predictions (such as top-5L and top-3L) than for small number of contact predictions such as top-L/2 and top-L/5. The results and analysis demonstrate that COMTOP can significantly improve the performance of the individual methods; therefore, COMTOP is more robust against different types of test sets. COMTOP also showed better/comparable predictions when compared with the state-of-the-art predictors.

Protein Structure Prediction: Conventional and Deep Learning Perspectives

Protein structure prediction is a way to bridge the sequence-structure gap, one of the main challenges in computational biology and chemistry. Predicting any protein's accurate structure is of paramount importance for the scientific community, as these structures govern their function. Moreover, this is one of the complicated optimization problems that computational biologists have ever faced. Experimental protein structure determination methods include X-ray crystallography, Nuclear Magnetic Resonance Spectroscopy and Electron Microscopy. All of these are tedious and time-consuming procedures that require expertise. To make the process less cumbersome, scientists use predictive tools as part of computational methods, using data consolidated in the protein repositories. In recent years, machine learning approaches have raised the interest of the structure prediction community. Most of the machine learning approaches for protein structure prediction are centred on co-evolution based methods. The accuracy of these approaches depends on the number of homologous protein sequences available in the databases. The prediction problem becomes challenging for many proteins, especially those without enough sequence homologs. Deep learning methods allow for the extraction of intricate features from protein sequence data without making any intuitions. Accurately predicted protein structures are employed for drug discovery, antibody designs, understanding protein-protein interactions, and interactions with other molecules. This article provides a review of conventional and deep learning approaches in protein structure prediction. We conclude this review by outlining a few publicly available datasets and deep learning architectures currently employed for protein structure prediction tasks.

Enhancing fragment-based protein structure prediction by customising fragment cardinality according to local secondary structure

BACKGROUND

Whenever suitable template structures are not available, usage of fragment-based protein structure prediction becomes the only practical alternative as pure ab initio techniques require massive computational resources even for very small proteins. However, inaccuracy of their energy functions and their stochastic nature imposes generation of a large number of decoys to explore adequately the solution space, limiting their usage to small proteins. Taking advantage of the uneven complexity of the sequence-structure relationship of short fragments, we adjusted the fragment insertion process by customising the number of available fragment templates according to the expected complexity of the predicted local secondary structure. Whereas the number of fragments is kept to its default value for coil regions, important and dramatic reductions are proposed for beta sheet and alpha helical regions, respectively.

RESULTS

The evaluation of our fragment selection approach was conducted using an enhanced version of the popular Rosetta fragment-based protein structure prediction tool. It was modified so that the number of fragment candidates used in Rosetta could be adjusted based on the local secondary structure. Compared to Rosetta's standard predictions, our strategy delivered improved first models, + 24% and + 6% in terms of GDT, when using 2000 and 20,000 decoys, respectively, while reducing significantly the number of fragment candidates. Furthermore, our enhanced version of Rosetta is able to deliver with 2000 decoys a performance equivalent to that produced by standard Rosetta while using 20,000 decoys. We hypothesise that, as the fragment insertion process focuses on the most challenging regions, such as coils, fewer decoys are needed to explore satisfactorily conformation spaces.

CONCLUSIONS

Taking advantage of the high accuracy of sequence-based secondary structure predictions, we showed the value of that information to customise the number of candidates used during the fragment insertion process of fragment-based protein structure prediction. Experimentations conducted using standard Rosetta showed that, when using the recommended number of decoys, i.e. 20,000, our strategy produces better results. Alternatively, similar results can be achieved using only 2000 decoys. Consequently, we recommend the adoption of this strategy to either improve significantly model quality or reduce processing times by a factor 10.

Be positive: optimizing pramlintide from microcanonical analysis of amylin isoforms

Amylin, or human islet amyloid polypeptide (hIAPP), is a 37-residue hormone synergistic to insulin and co-secreted with it by β-cells in the pancreas. The deposition of its cytotoxic amyloid fibrils is strongly related to the progression of Type II diabetes (T2D) and islet graft failures. Notably, isoforms from some mammalian species, such as rats (rIAPP) and porcine (pIAPP), present a few key mutations preventing aggregation. This has lead to biotechnological development of drugs for adjunct therapies of T2D, such as pramlintide, a variant of hIAPP inspired by rIAPP whose proline substitutions have β-strand fibril-breaking properties. Ideally, such a drug should be formulated with insulin and co-administered, but this has been prevented by a poor solubility profile at the appropriate pH. Hopefully, this could be improved with appropriate point mutations, increasing the molecular net charge. Despite experimental progress, preliminary screening during rational drug design can greatly benefit from thermodynamic insight derived from molecular simulations. So we introduce microcanonical thermostatistics analysis of multicanonical (MUCA) simulations of wild-type amylin isoforms as a systematic assessment of protein thermostability. As a consequence of this comprehensive investigation, the most suitable single-point mutations able to optimize pramlintide are located among the wild-type amylin isoforms. In particular, we find that aggregation inhibition and increased solubility are inherited by pramlintide through further S20R substitution typical of pIAPP. Thus, we provide a consistent thermostatistical methodology to aid the design of improved adjunct therapies for T2D according to current clinical knowledge.

COMSAT: Residue contact prediction of transmembrane proteins based on support vector machines and mixed integer linear programming

In this article, we present COMSAT, a hybrid framework for residue contact prediction of transmembrane (TM) proteins, integrating a support vector machine (SVM) method and a mixed integer linear programming (MILP) method. COMSAT consists of two modules: COMSAT_SVM which is trained mainly on position-specific scoring matrix features, and COMSAT_MILP which is an ab initio method based on optimization models. Contacts predicted by the SVM model are ranked by SVM confidence scores, and a threshold is trained to improve the reliability of the predicted contacts. For TM proteins with no contacts above the threshold, COMSAT_MILP is used. The proposed hybrid contact prediction scheme was tested on two independent TM protein sets based on the contact definition of 14 Å between Cα-Cα atoms. First, using a rigorous leave-one-protein-out cross validation on the training set of 90 TM proteins, an accuracy of 66.8%, a coverage of 12.3%, a specificity of 99.3% and a Matthews' correlation coefficient (MCC) of 0.184 were obtained for residue pairs that are at least six amino acids apart. Second, when tested on a test set of 87 TM proteins, the proposed method showed a prediction accuracy of 64.5%, a coverage of 5.3%, a specificity of 99.4% and a MCC of 0.106. COMSAT shows satisfactory results when compared with 12 other state-of-the-art predictors, and is more robust in terms of prediction accuracy as the length and complexity of TM protein increase. COMSAT is freely accessible at http://hpcc.siat.ac.cn/COMSAT/.

Improved contact predictions using the recognition of protein like contact patterns

Given sufficient large protein families, and using a global statistical inference approach, it is possible to obtain sufficient accuracy in protein residue contact predictions to predict the structure of many proteins. However, these approaches do not consider the fact that the contacts in a protein are neither randomly, nor independently distributed, but actually follow precise rules governed by the structure of the protein and thus are interdependent. Here, we present PconsC2, a novel method that uses a deep learning approach to identify protein-like contact patterns to improve contact predictions. A substantial enhancement can be seen for all contacts independently on the number of aligned sequences, residue separation or secondary structure type, but is largest for β-sheet containing proteins. In addition to being superior to earlier methods based on statistical inferences, in comparison to state of the art methods using machine learning, PconsC2 is superior for families with more than 100 effective sequence homologs. The improved contact prediction enables improved structure prediction.

Here, we introduce a novel protein contact prediction method PconsC2 that, to the best of our knowledge, outperforms earlier methods. PconsC2 is based on our earlier method, PconsC, as it utilizes the same set of contact predictions from plmDCA and PSICOV. However, in contrast to PconsC, where each residue pair is analysed independently, the initial predictions are analysed in context of neighbouring residue pairs using a deep learning approach, inspired by earlier work. We find that for each layer the deep learning procedure improves the predictions. At the end, after five layers of deep learning and inclusion of a few extra features provides the best performance. An improvement can be seen for all types of proteins, independent on length, number of homologous sequences and structural class. However, the improvement is largest for β-sheet containing proteins. Most importantly the improvement brings for the first time sufficiently accurate predictions to some protein families with less than 1000 homologous sequences. PconsC2 outperforms as well state of the art machine learning based predictors for protein families larger than 100 effective sequences. PconsC2 is licensed under the GNU General Public License v3 and freely available from http://c2.pcons.net/.

Forcefield_PTM: Ab Initio Charge and AMBER Forcefield Parameters for Frequently Occurring Post-Translational Modifications

In this work, we introduce Forcefield_PTM, a set of AMBER forcefield parameters consistent with ff03 for 32 common post-translational modifications. Partial charges were calculated through ab initio calculations and a two-stage RESP-fitting procedure in an ether-like implicit solvent environment. The charges were found to be generally consistent with others previously reported for phosphorylated amino acids, and trimethyllysine, using different parameterization methods. Pairs of modified and their corresponding unmodified structures were curated from the PDB for both single and multiple modifications. Background structural similarity was assessed in the context of secondary and tertiary structures from the global dataset. Next, the charges derived for Forcefield_PTM were tested on a macroscopic scale using unrestrained all-atom Langevin molecular dynamics simulations in AMBER for 34 (17 pairs of modified/unmodified) systems in implicit solvent. Assessment was performed in the context of secondary structure preservation, stability in energies, and correlations between the modified and unmodified structure trajectories on the aggregate. As an illustration of their utility, the parameters were used to compare the structural stability of the phosphorylated and dephosphorylated forms of OdhI. Microscopic comparisons between quantum and AMBER single point energies along key χ torsions on several PTMs were performed and corrections to improve their agreement in terms of mean squared errors and squared correlation coefficients were parameterized. This forcefield for post-translational modifications in condensed-phase simulations can be applied to a number of biologically relevant and timely applications including protein structure prediction, protein and peptide design, docking, and to study the effect of PTMs on folding and dynamics. We make the derived parameters and an associated interactive webtool capable of performing post-translational modifications on proteins using Forcefield_PTM available at http://selene.princeton.edu/FFPTM.

Princeton_TIGRESS: protein geometry refinement using simulations and support vector machines

Protein structure refinement aims to perform a set of operations given a predicted structure to improve model quality and accuracy with respect to the native in a blind fashion. Despite the numerous computational approaches to the protein refinement problem reported in the previous three CASPs, an overwhelming majority of methods degrade models rather than improve them. We initially developed a method tested using blind predictions during CASP10 which was officially ranked in 5th place among all methods in the refinement category. Here, we present Princeton_TIGRESS, which when benchmarked on all CASP 7,8,9, and 10 refinement targets, simultaneously increased GDT_TS 76% of the time with an average improvement of 0.83 GDT_TS points per structure. The method was additionally benchmarked on models produced by top performing three-dimensional structure prediction servers during CASP10. The robustness of the Princeton_TIGRESS protocol was also tested for different random seeds. We make the Princeton_TIGRESS refinement protocol freely available as a web server at http://atlas.princeton.edu/refinement. Using this protocol, one can consistently refine a prediction to help bridge the gap between a predicted structure and the actual native structure.

Protein folding and de novo protein design for biotechnological applications

In the postgenomic era, the medical/biological fields are advancing faster than ever. However, before the power of full-genome sequencing can be fully realized, the connection between amino acid sequence and protein structure, known as the protein folding problem, needs to be elucidated. The protein folding problem remains elusive, with significant difficulties still arising when modeling amino acid sequences lacking an identifiable template. Understanding protein folding will allow for unforeseen advances in protein design; often referred to as the inverse protein folding problem. Despite challenges in protein folding, de novo protein design has recently demonstrated significant success via computational techniques. We review advances and challenges in protein structure prediction and de novo protein design, and highlight their interplay in successful biotechnological applications.

Protein WISDOM: a workbench for in silico de novo design of biomolecules

The aim of de novo protein design is to find the amino acid sequences that will fold into a desired 3-dimensional structure with improvements in specific properties, such as binding affinity, agonist or antagonist behavior, or stability, relative to the native sequence. Protein design lies at the center of current advances drug design and discovery. Not only does protein design provide predictions for potentially useful drug targets, but it also enhances our understanding of the protein folding process and protein-protein interactions. Experimental methods such as directed evolution have shown success in protein design. However, such methods are restricted by the limited sequence space that can be searched tractably. In contrast, computational design strategies allow for the screening of a much larger set of sequences covering a wide variety of properties and functionality. We have developed a range of computational de novo protein design methods capable of tackling several important areas of protein design. These include the design of monomeric proteins for increased stability and complexes for increased binding affinity. To disseminate these methods for broader use we present Protein WISDOM (http://www.proteinwisdom.org), a tool that provides automated methods for a variety of protein design problems. Structural templates are submitted to initialize the design process. The first stage of design is an optimization sequence selection stage that aims at improving stability through minimization of potential energy in the sequence space. Selected sequences are then run through a fold specificity stage and a binding affinity stage. A rank-ordered list of the sequences for each step of the process, along with relevant designed structures, provides the user with a comprehensive quantitative assessment of the design. Here we provide the details of each design method, as well as several notable experimental successes attained through the use of the methods.

β-sheet topology prediction with high precision and recall for β and mixed α/β proteins

The prediction of the correct -sheet topology for pure and mixed proteins is a critical intermediate step toward the three dimensional protein structure prediction. The predicted beta sheet topology provides distance constraints between sequentially separated residues, which reduces the three dimensional search space for a protein structure prediction algorithm. Here, we present a novel mixed integer linear optimization based framework for the prediction of -sheet topology in and mixed proteins. The objective is to maximize the total strand-to-strand contact potential of the protein. A large number of physical constraints are applied to provide biologically meaningful topology results. The formulation permits the creation of a rank-ordered list of preferred -sheet arrangements. Finally, the generated topologies are re-ranked using a fully atomistic approach involving torsion angle dynamics and clustering. For a large, non-redundant data set of 2102 and mixed proteins with at least 3 strands taken from the PDB, the proposed approach provides the top 5 solutions with average precision and recall greater than 78%. Consistent results are obtained in the -sheet topology prediction for blind targets provided during the CASP8 and CASP9 experiments, as well as for actual and predicted secondary structures. The -sheet topology prediction algorithm, BeST, is available to the scientific community at http://selene.princeton.edu/BeST/.