Single-sequence-based prediction of protein secondary structures and solvent accessibility by deep whole-sequence learning

Can large language models predict antimicrobial peptide activity and toxicity?

Antimicrobial peptides (AMPs) are naturally occurring or designed peptides up to a few tens of amino acids which may help address the antimicrobial resistance crisis. However, their clinical development is limited by toxicity to human cells, a parameter which is very difficult to control. Given the similarity between peptide sequences and words, large language models (LLMs) might be able to predict AMP activity and toxicity. To test this hypothesis, we fine-tuned LLMs using data from the Database of Antimicrobial Activity and Structure of Peptides (DBAASP). GPT-3 performed well but not reproducibly for activity prediction and hemolysis, taken as a proxy for toxicity. The later GPT-3.5 performed more poorly and was surpassed by recurrent neural networks (RNN) trained on sequence-activity data or support vector machines (SVM) trained on MAP4C molecular fingerprint-activity data. These simpler models are therefore recommended, although the rapid evolution of LLMs warrants future re-evaluation of their prediction abilities.

Random, de novo, and conserved proteins: How structure and disorder predictors perform differently

Understanding the emergence and structural characteristics of de novo and random proteins is crucial for unraveling protein evolution and designing novel enzymes. However, experimental determination of their structures remains challenging. Recent advancements in protein structure prediction, particularly with AlphaFold2 (AF2), have expanded our knowledge of protein structures, but their applicability to de novo and random proteins is unclear. In this study, we investigate the structural predictions and confidence scores of AF2 and protein language model-based predictor ESMFold for de novo and conserved proteins from Drosophila and a dataset of comparable random proteins. We find that the structural predictions for de novo and random proteins differ significantly from conserved proteins. Interestingly, a positive correlation between disorder and confidence scores (pLDDT) is observed for de novo and random proteins, in contrast to the negative correlation observed for conserved proteins. Furthermore, the performance of structure predictors for de novo and random proteins is hampered by the lack of sequence identity. We also observe fluctuating median predicted disorder among different sequence length quartiles for random proteins, suggesting an influence of sequence length on disorder predictions. In conclusion, while structure predictors provide initial insights into the structural composition of de novo and random proteins, their accuracy and applicability to such proteins remain limited. Experimental determination of their structures is necessary for a comprehensive understanding. The positive correlation between disorder and pLDDT could imply a potential for conditional folding and transient binding interactions of de novo and random proteins.

PARP inhibitors suppress tumours via centrosome error-induced senescence independent of DNA damage response

BACKGROUND

Poly(ADP-ribose) polymerase (PARP) inhibitors have emerged as promising chemotherapeutic drugs primarily against BRCA1/2-associated tumours, known as synthetic lethality. However, recent clinical trials reported patients' survival benefits from PARP inhibitor treatments, irrelevant to homologous recombination deficiency. Therefore, revealing the therapeutic mechanism of PARP inhibitors beyond DNA damage repair is urgently needed, which can facilitate precision medicine.

METHODS

A CRISPR-based knock-in technology was used to establish stable BRCA1 mutant cancer cells. The effects of PARP inhibitors on BRCA1 mutant cancer cells were evaluated by biochemical and cell biological experiments. Finally, we validated its in vivo effects in xenograft and patient-derived xenograft (PDX) tumour mice.

FINDINGS

In this study, we uncovered that the majority of clinical BRCA1 mutations in breast cancers were in and near the middle of the gene, rather than in essential regions for DNA damage repair. Representative mutations such as R1085I and E1222Q caused transient extra spindle poles during mitosis in cancer cells. PAR, which is synthesized by PARP2 but not PARP1 at mitotic centrosomes, clustered these transient extra poles, independent of DNA damage response. Common PARP inhibitors could effectively suppress PARP2-synthesized PAR and induce cell senescence by abrogating the correction of mitotic extra-pole error.

INTERPRETATION

Our findings uncover an alternative mechanism by which PARP inhibitors efficiently suppress tumours, thereby pointing to a potential new therapeutic strategy for centrosome error-related tumours.

FUNDING

Funded by National Natural Science Foundation of China (NSFC) (T2225006, 82272948, 82103106), Beijing Municipal Natural Science Foundation (Key program Z220011), and the National Clinical Key Specialty Construction Program, P. R. China (2023).

Novel antimicrobial peptides against Cutibacterium acnes designed by deep learning

The increasing prevalence of antibiotic resistance in Cutibacterium acnes (C. acnes) requires the search for alternative therapeutic strategies. Antimicrobial peptides (AMPs) offer a promising avenue for the development of new treatments targeting C. acnes. In this study, to design peptides with the specific inhibitory activity against C. acnes, we employed a deep learning pipeline with generators and classifiers, using transfer learning and pretrained protein embeddings, trained on publicly available data. To enhance the training data specific to C. acnes inhibition, we constructed a phylogenetic tree. A panel of 42 novel generated linear peptides was then synthesized and experimentally evaluated for their antimicrobial selectivity and activity. Five of them demonstrated their high potency and selectivity against C. acnes with MIC of 2-4 µg/mL. Our findings highlight the potential of these designed peptides as promising candidates for anti-acne therapeutics and demonstrate the power of computational approaches for the rational design of targeted antimicrobial peptides.

Naive Prediction of Protein Backbone Phi and Psi Dihedral Angles Using Deep Learning

Protein structure prediction represents a significant challenge in the field of bioinformatics, with the prediction of protein structures using backbone dihedral angles recently achieving significant progress due to the rise of deep neural network research. However, there is a trend in protein structure prediction research to employ increasingly complex neural networks and contributions from multiple models. This study, on the other hand, explores how a single model transparently behaves using sequence data only and what can be expected from the predicted angles. To this end, the current paper presents data acquisition, deep learning model definition, and training toward the final protein backbone angle prediction. The method applies a simple fully connected neural network (FCNN) model that takes only the primary structure of the protein with a sliding window of size 21 as input to predict protein backbone ϕ and ψ dihedral angles. Despite its simplicity, the model shows surprising accuracy for the ϕ angle prediction and somewhat lower accuracy for the ψ angle prediction. Moreover, this study demonstrates that protein secondary structure prediction is also possible with simple neural networks that take in only the protein amino-acid residue sequence, but more complex models are required for higher accuracies.

Protein Structure Prediction: Challenges, Advances, and the Shift of Research Paradigms

Protein structure prediction is an interdisciplinary research topic that has attracted researchers from multiple fields, including biochemistry, medicine, physics, mathematics, and computer science. These researchers adopt various research paradigms to attack the same structure prediction problem: biochemists and physicists attempt to reveal the principles governing protein folding; mathematicians, especially statisticians, usually start from assuming a probability distribution of protein structures given a target sequence and then find the most likely structure, while computer scientists formulate protein structure prediction as an optimization problem - finding the structural conformation with the lowest energy or minimizing the difference between predicted structure and native structure. These research paradigms fall into the two statistical modeling cultures proposed by Leo Breiman, namely, data modeling and algorithmic modeling. Recently, we have also witnessed the great success of deep learning in protein structure prediction. In this review, we present a survey of the efforts for protein structure prediction. We compare the research paradigms adopted by researchers from different fields, with an emphasis on the shift of research paradigms in the era of deep learning. In short, the algorithmic modeling techniques, especially deep neural networks, have considerably improved the accuracy of protein structure prediction; however, theories interpreting the neural networks and knowledge on protein folding are still highly desired.

Efficient mapping of RNA-binding residues in RNA-binding proteins using local sequence features of binding site residues in protein-RNA complexes

Protein-RNA interactions play vital roles in plethora of biological processes such as regulation of gene expression, protein synthesis, mRNA processing and biogenesis. Identification of RNA-binding residues (RBRs) in proteins is essential to understand RNA-mediated protein functioning, to perform site-directed mutagenesis and to develop novel targeted drug therapies. Moreover, the extensive gap between sequence and structural data restricts the identification of binding sites in unsolved structures. However, efficient use of computational methods demanding only sequence to identify binding residues can bridge this huge sequence-structure gap. In this study, we have extensively studied protein-RNA interface in known RNA-binding proteins (RBPs). We find that the interface is highly enriched in basic and polar residues with Gly being the most common interface neighbor. We investigated several amino acid features and developed a method to predict putative RBRs from amino acid sequence. We have implemented balanced random forest (BRF) classifier with local residue features of protein sequences for prediction. With 5-fold cross-validations, the sequence pattern derived dipeptide composition based BRF model (DCP-BRF) resulted in an accuracy of 87.9%, specificity of 88.8%, sensitivity of 82.2%, Mathew's correlation coefficient of 0.60 and AUC of 0.93, performing better than few existing methods. We further validated our prediction model on known human RBPs through RBR prediction and could map ~54% of them. Further, knowledge of binding site preferences obtained from computational predictions combined with experimental validations of potential RNA binding sites can enhance our understanding of protein-RNA interactions. This may serve to accelerate investigations on functional roles of many novel RBPs.

AttSec: protein secondary structure prediction by capturing local patterns from attention map

BACKGROUND

Protein secondary structures that link simple 1D sequences to complex 3D structures can be used as good features for describing the local properties of protein, but also can serve as key features for predicting the complex 3D structures of protein. Thus, it is very important to accurately predict the secondary structure of the protein, which contains a local structural property assigned by the pattern of hydrogen bonds formed between amino acids. In this study, we accurately predict protein secondary structure by capturing the local patterns of protein. For this objective, we present a novel prediction model, AttSec, based on transformer architecture. In particular, AttSec extracts self-attention maps corresponding to pairwise features between amino acid embeddings and passes them through 2D convolution blocks to capture local patterns. In addition, instead of using additional evolutionary information, it uses protein embedding as an input, which is generated by a language model.

RESULTS

For the ProteinNet DSSP8 dataset, our model showed 11.8% better performance on the entire evaluation datasets compared with other no-evolutionary-information-based models. For the NetSurfP-2.0 DSSP8 dataset, it showed 1.2% better performance on average. There was an average performance improvement of 9.0% for the ProteinNet DSSP3 dataset and an average of 0.7% for the NetSurfP-2.0 DSSP3 dataset.

CONCLUSION

We accurately predict protein secondary structure by capturing the local patterns of protein. For this objective, we present a novel prediction model, AttSec, based on transformer architecture. Although there was no dramatic accuracy improvement compared with other models, the improvement on DSSP8 was greater than that on DSSP3. This result implies that using our proposed pairwise feature could have a remarkable effect for several challenging tasks that require finely subdivided classification. Github package URL is https://github.com/youjin-DDAI/AttSec .

Personal Precise Force Field for Intrinsically Disordered and Ordered Proteins Based on Deep Learning

Intrinsically disordered proteins (IDPs) are proteins without a fixed three-dimensional (3D) structure under physiological conditions and are associated with Parkinson's disease, Alzheimer's disease, cancer, cardiovascular disease, amyloidosis, diabetes, and other diseases. Experimental methods can hardly capture the ensemble of diverse conformations for IDPs. Molecular dynamics (MD) simulations can sample continuous conformations that might provide a valuable complement to experimental data. However, the accuracy of MD simulations depends on the quality of force field. In particular, the evolutionary conservation and coevolution of IDPs introduce that current force fields could not precisely reproduce the conformation of IDPs. In order to improve the performance of force field, deep learning and reweighting methods were used to automatically generate personal force field parameters for intrinsically disordered and ordered proteins. At first, the deep learning method predicted more accuracy φ/ψ dihedral of residue than the previous method. Then, reweighting optimized the personal force field parameters for each residue. Finally, typical representative systems such as IDPs, structure protein, and fast-folding protein were used to evaluate this force field. The results indicate that two personal force field parameters (named PPFF1 and PPFF1_af2) could better reproduce the experimental observables than ff03CMAP force field. In summary, this strategy will provide feasibility for the development of precise personal force fields.

Machine learning approaches demonstrate that protein structures carry information about their genetic coding

Synonymous codons translate into the same amino acid. Although the identity of synonymous codons is often considered inconsequential to the final protein structure, there is mounting evidence for an association between the two. Our study examined this association using regression and classification models, finding that codon sequences predict protein backbone dihedral angles with a lower error than amino acid sequences, and that models trained with true dihedral angles have better classification of synonymous codons given structural information than models trained with random dihedral angles. Using this classification approach, we investigated local codon-codon dependencies and tested whether synonymous codon identity can be predicted more accurately from codon context than amino acid context alone, and most specifically which codon context position carries the most predictive power.

Multi-model predictive analysis of RNA solvent accessibility based on modified residual attention mechanism

Predicting RNA solvent accessibility using only primary sequence data can be regarded as sequence-based prediction work. Currently, the established studies for sequence-based RNA solvent accessibility prediction are limited due to the available number of datasets and black box prediction. To improve these issues, we first expanded the available RNA structures and then developed a sequence-based model using modified attention layers with different receptive fields to conform to the stem-loop structure of RNA chains. We measured the improvement with an extended dataset and further explored the model's interpretability by analysing the model structures, attention values and hyperparameters. Finally, we found that the developed model regarded the pieces of a sequence as templates during the training process. This work will be helpful for researchers who would like to build RNA attribute prediction models using deep learning in the future.

Deep learning for protein secondary structure prediction: Pre and post-AlphaFold

This paper aims to provide a comprehensive review of the trends and challenges of deep neural networks for protein secondary structure prediction (PSSP). In recent years, deep neural networks have become the primary method for protein secondary structure prediction. Previous studies showed that deep neural networks had uplifted the accuracy of three-state secondary structure prediction to more than 80%. Favored deep learning methods, such as convolutional neural networks, recurrent neural networks, inception networks, and graph neural networks, have been implemented in protein secondary structure prediction. Methods adapted from natural language processing (NLP) and computer vision are also employed, including attention mechanism, ResNet, and U-shape networks. In the post-AlphaFold era, PSSP studies focus on different objectives, such as enhancing the quality of evolutionary information and exploiting protein language models as the PSSP input. The recent trend to utilize pre-trained language models as input features for secondary structure prediction provides a new direction for PSSP studies. Moreover, the state-of-the-art accuracy achieved by previous PSSP models is still below its theoretical limit. There are still rooms for improvement to be made in the field.

To Improve Prediction of Binding Residues With DNA, RNA, Carbohydrate, and Peptide Via Multi-Task Deep Neural Networks

MOTIVATION

The interactions of proteins with DNA, RNA, peptide, and carbohydrate play key roles in various biological processes. The studies of uncharacterized protein-molecules interactions could be aided by accurate predictions of residues that bind with partner molecules. However, the existing methods for predicting binding residues on proteins remain of relatively low accuracies due to the limited number of complex structures in databases. As different types of molecules partially share chemical mechanisms, the predictions for each molecular type should benefit from the binding information with other molecule types.

RESULTS

In this study, we employed a multiple task deep learning strategy to develop a new sequence-based method for simultaneously predicting binding residues/sites with multiple important molecule types named MTDsite. By combining four training sets for DNA, RNA, peptide, and carbohydrate-binding proteins, our method yielded accurate and robust predictions with AUC values of 0.852, 0836, 0.758, and 0.776 on their respective independent test sets, which are 0.52 to 6.6% better than other state-of-the-art methods. To my best knowledge, this is the first method using multi-task framework to predict multiple molecular binding sites simultaneously.

Similar protein segments shared between domains of different evolutionary lineages

The emergence of novel proteins, beyond these that can be readily made by duplication and recombination of preexisting domains, is elusive. De novo emergence from random sequences is unlikely because the vast majority of random chains would not even fold, let alone function. An alternative explanation is that novel proteins emerge by duplication and fusion of pre‐existing polypeptide segments. In this case, traces of such ancient events may remain within contemporary proteins in the form of reused segments. Together with the late Dan Tawfik, we detected such similar segments, far shorter than intact protein domains, which are found in different environments. The detection of these, “bridging themes,” was based on a unique search strategy, where in addition to searching for similarity of shared fragments, so‐called “themes,” we also explicitly searched for cases in which the sequence segments before and after the theme are dissimilar (both in sequence and structure). Here, using a similar strategy, we further expanded the search and discovered almost 500 additional “bridging themes,” linking domains that are often from ancient folds. The themes, of 20 residues or more (average 53), do not retain their structure despite sharing 37% sequence identity on average. Indeed, conformation flexibility may confer an evolutionary advantage, in that it fits in multiple environments. We elaborate on two interesting themes, shared between Rossmann/Trefoil‐Plexin‐like domains and a β‐propeller‐like domain.

PROST: AlphaFold2-aware Sequence-Based Predictor to Estimate Protein Stability Changes upon Missense Mutations

An essential step in engineering proteins and understanding disease-causing missense mutations is to accurately model protein stability changes when such mutations occur. Here, we developed a new sequence-based predictor for the protein stability (PROST) change (Gibb's free energy change, ΔΔG) upon a single-point missense mutation. PROST extracts multiple descriptors from the most promising sequence-based predictors, such as BoostDDG, SAAFEC-SEQ, and DDGun. RPOST also extracts descriptors from iFeature and AlphaFold2. The extracted descriptors include sequence-based features, physicochemical properties, evolutionary information, evolutionary-based physicochemical properties, and predicted structural features. The PROST predictor is a weighted average ensemble model based on extreme gradient boosting (XGBoost) decision trees and an extra-trees regressor; PROST is trained on both direct and hypothetical reverse mutations using the S5294 (S2647 direct mutations + S2647 inverse mutations). The parameters for the PROST model are optimized using grid searching with 5-fold cross-validation, and feature importance analysis unveils the most relevant features. The performance of PROST is evaluated in a blinded manner, employing nine distinct data sets and existing state-of-the-art sequence-based and structure-based predictors. This method consistently performs well on frataxin, S217, S349, Ssym, S669, Myoglobin, and CAGI5 data sets in blind tests and similarly to the state-of-the-art predictors for p53 and S276 data sets. When the performance of PROST is compared with the latest predictors such as BoostDDG, SAAFEC-SEQ, ACDC-NN-seq, and DDGun, PROST dominates these predictors. A case study of mutation scanning of the frataxin protein for nine wild-type residues demonstrates the utility of PROST. Taken together, these findings indicate that PROST is a well-suited predictor when no protein structural information is available. The source code of PROST, data sets, examples, and pretrained models along with how to use PROST are available at https://github.com/ShahidIqb/PROST and https://prost.erc.monash.edu/seq.

Machine Learning Guided Discovery of Non-Hemolytic Membrane Disruptive Anti-Cancer Peptides

Most antimicrobial peptides (AMPs) and anticancer peptides (ACPs) fold into membrane disruptive cationic amphiphilic α‐helices, many of which are however also unpredictably hemolytic and toxic. Here we exploited the ability of recurrent neural networks (RNN) to distinguish active from inactive and non‐hemolytic from hemolytic AMPs and ACPs to discover new non‐hemolytic ACPs. Our discovery pipeline involved: 1) sequence generation using either a generative RNN or a genetic algorithm, 2) RNN classification for activity and hemolysis, 3) selection for sequence novelty, helicity and amphiphilicity, and 4) synthesis and testing. Experimental evaluation of thirty‐three peptides resulted in eleven active ACPs, four of which were non‐hemolytic, with properties resembling those of the natural ACP lasioglossin III. These experiments show the first example of direct machine learning guided discovery of non‐hemolytic ACPs.

ContactPFP: Protein Function Prediction Using Predicted Contact Information

Computational function prediction is one of the most important problems in bioinformatics as elucidating the function of genes is a central task in molecular biology and genomics. Most of the existing function prediction methods use protein sequences as the primary source of input information because the sequence is the most available information for query proteins. There are attempts to consider other attributes of query proteins. Among these attributes, the three-dimensional (3D) structure of proteins is known to be very useful in identifying the evolutionary relationship of proteins, from which functional similarity can be inferred. Here, we report a novel protein function prediction method, ContactPFP, which uses predicted residue-residue contact maps as input structural features of query proteins. Although 3D structure information is known to be useful, it has not been routinely used in function prediction because the 3D structure is not experimentally determined for many proteins. In ContactPFP, we overcome this limitation by using residue-residue contact prediction, which has become increasingly accurate due to rapid development in the protein structure prediction field. ContactPFP takes a query protein sequence as input and uses predicted residue-residue contact as a proxy for the 3D protein structure. To characterize how predicted contacts contribute to function prediction accuracy, we compared the performance of ContactPFP with several well-established sequence-based function prediction methods. The comparative study revealed the advantages and weaknesses of ContactPFP compared to contemporary sequence-based methods. There were many cases where it showed higher prediction accuracy. We examined factors that affected the accuracy of ContactPFP using several illustrative cases that highlight the strength of our method.

Modern and prebiotic amino acids support distinct structural profiles in proteins

The earliest proteins had to rely on amino acids available on early Earth before the biosynthetic pathways for more complex amino acids evolved. In extant proteins, a significant fraction of the 'late' amino acids (such as Arg, Lys, His, Cys, Trp and Tyr) belong to essential catalytic and structure-stabilizing residues. How (or if) early proteins could sustain an early biosphere has been a major puzzle. Here, we analysed two combinatorial protein libraries representing proxies of the available sequence space at two different evolutionary stages. The first is composed of the entire alphabet of 20 amino acids while the second one consists of only 10 residues (ASDGLIPTEV) representing a consensus view of plausibly available amino acids through prebiotic chemistry. We show that compact conformations resistant to proteolysis are surprisingly similarly abundant in both libraries. In addition, the early alphabet proteins are inherently more soluble and refoldable, independent of the general Hsp70 chaperone activity. By contrast, chaperones significantly increase the otherwise poor solubility of the modern alphabet proteins suggesting their coevolution with the amino acid repertoire. Our work indicates that while both early and modern amino acids are predisposed to supporting protein structure, they do so with different biophysical properties and via different mechanisms.

BIAPSS: A Comprehensive Physicochemical Analyzer of Proteins Undergoing Liquid-Liquid Phase Separation

The liquid-liquid phase separation (LLPS) of biomolecules is a phenomenon which is nowadays recognized as the driving force for the biogenesis of numerous functional membraneless organelles and cellular bodies. The interplay between the protein primary sequence and phase separation remains poorly understood, despite intensive research. To uncover the sequence-encoded signals of protein capable of undergoing LLPS, we developed a novel web platform named BIAPSS (Bioinformatics Analysis of LLPS Sequences). This web server provides on-the-fly analysis, visualization, and interpretation of the physicochemical and structural features for the superset of curated LLPS proteins.

Reaching alignment-profile-based accuracy in predicting protein secondary and tertiary structural properties without alignment

Protein language models have emerged as an alternative to multiple sequence alignment for enriching sequence information and improving downstream prediction tasks such as biophysical, structural, and functional properties. Here we show that a method called SPOT-1D-LM combines traditional one-hot encoding with the embeddings from two different language models (ProtTrans and ESM-1b) for the input and yields a leap in accuracy over single-sequence-based techniques in predicting protein 1D secondary and tertiary structural properties, including backbone torsion angles, solvent accessibility and contact numbers for all six test sets (TEST2018, TEST2020, Neff1-2020, CASP12-FM, CASP13-FM and CASP14-FM). More significantly, it has a performance comparable to profile-based methods for those proteins with homologous sequences. For example, the accuracy for three-state secondary structure (SS3) prediction for TEST2018 and TEST2020 proteins are 86.7% and 79.8% by SPOT-1D-LM, compared to 74.3% and 73.4% by the single-sequence-based method SPOT-1D-Single and 86.2% and 80.5% by the profile-based method SPOT-1D, respectively. For proteins without homologous sequences (Neff1-2020) SS3 is 80.41% by SPOT-1D-LM which is 3.8% and 8.3% higher than SPOT-1D-Single and SPOT-1D, respectively. SPOT-1D-LM is expected to be useful for genome-wide analysis given its fast performance. Moreover, high-accuracy prediction of both secondary and tertiary structural properties such as backbone angles and solvent accessibility without sequence alignment suggests that highly accurate prediction of protein structures may be made without homologous sequences, the remaining obstacle in the post AlphaFold2 era.

Complementarity of the residue-level protein function and structure predictions in human proteins

Sequence-based predictors of the residue-level protein function and structure cover a broad spectrum of characteristics including intrinsic disorder, secondary structure, solvent accessibility and binding to nucleic acids. They were catalogued and evaluated in numerous surveys and assessments. However, methods focusing on a given characteristic are studied separately from predictors of other characteristics, while they are typically used on the same proteins. We fill this void by studying complementarity of a representative collection of methods that target different predictions using a large, taxonomically consistent, and low similarity dataset of human proteins. First, we bridge the gap between the communities that develop structure-trained vs. disorder-trained predictors of binding residues. Motivated by a recent study of the protein-binding residue predictions, we empirically find that combining the structure-trained and disorder-trained predictors of the DNA-binding and RNA-binding residues leads to substantial improvements in predictive quality. Second, we investigate whether diverse predictors generate results that accurately reproduce relations between secondary structure, solvent accessibility, interaction sites, and intrinsic disorder that are present in the experimental data. Our empirical analysis concludes that predictions accurately reflect all combinations of these relations. Altogether, this study provides unique insights that support combining results produced by diverse residue-level predictors of protein function and structure.

proANP Metabolism Provides New Insights Into Sacubitril/Valsartan Mode of Action

BACKGROUND

Sacubitril/valsartan (S/V) treatment is associated with clinical benefits in patients with heart failure with reduced ejection fraction (HFrEF), but its mode of action remains elusive, although it involves the increase of ANP (atrial natriuretic peptide).

METHODS AND RESULTS

Using a cohort of 73 HFrEF patients treated with S/V and controls, we deciphered the proteolytic cascade that converts proANP into 4 vasoactive peptides, including ANP, which exert vasodilatory actions. We found that proANP processing is sequential and involved meprin B, ECE (endothelin-converting enzyme) 1, and ANPEP (aminopeptidase N). This processing is limited in HFrEF patients when compared with controls via the downregulation of proANP production, corin, and meprin B activities by miR-425 and miR1-3p, resulting in limited production of proANP-derived bioactive peptides. S/V restored or compensated proANP processing by downregulating miR-425 and miR1-3p beyond levels observed in controls, hence increasing levels of proANP-derived bioactive peptides and vasodilation. In contrast, S/V directly and indirectly partially inhibited ECE1 and ANPEP. Consequently, ECE1 partial inhibition resulted in a lower-than-expected increase in ET1 (endothelin 1), tilting the vasoactive balance toward vasodilation, possibly explaining the hypotensive action of S/V. Finally, we show that proANP glycosylation interferes with the midregional proANP assay-a clinical surrogate for proANP production, preventing any pathophysiological interpretation of the results. Finally, the analysis of S/V dose escalation with respect to baseline treatments suggests S/V-specific effects.

CONCLUSIONS

These findings offer mechanistic evidence to the natriuretic peptide-defective state in HFrEF, which is improved by S/V. These data also strongly suggest that S/V increases plasma ANP by multiple mechanisms that involve the indirect regulation of 2 microRNAs, besides its protection from NEP (neprilysin) cleavage. Altogether, these data provide new insights on HFrEF pathophysiology and the mode of action of S/V.

SPOT-Contact-LM: improving single-sequence-based prediction of protein contact map using a transformer language model

MOTIVATION

Accurate prediction of protein contact-map is essential for accurate protein structure and function prediction. As a result, many methods have been developed for protein contact map prediction. However, most methods rely on protein-sequence-evolutionary information, which may not exist for many proteins due to lack of naturally occurring homologous sequences. Moreover, generating evolutionary profiles is computationally intensive. Here, we developed a contact-map predictor utilizing the output of a pre-trained language model ESM-1b as an input along with a large training set and an ensemble of residual neural networks.

RESULTS

We showed that the proposed method makes a significant improvement over a single-sequence-based predictor SSCpred with 15% improvement in the F1-score for the independent CASP14-FM test set. It also outperforms evolutionary-profile-based methods trRosetta and SPOT-Contact with 48.7% and 48.5% respective improvement in the F1-score on the proteins without homologs (Neff = 1) in the independent SPOT-2018 set. The new method provides a much faster and reasonably accurate alternative to evolution-based methods, useful for large-scale prediction.

AVAILABILITY AND IMPLEMENTATION

Stand-alone-version of SPOT-Contact-LM is available at https://github.com/jas-preet/SPOT-Contact-Single. Direct prediction can also be made at https://sparks-lab.org/server/spot-contact-single. The datasets used in this research can also be downloaded from the GitHub.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Secondary structure specific simpler prediction models for protein backbone angles

Motivation

Protein backbone angle prediction has achieved significant accuracy improvement with the development of deep learning methods. Usually the same deep learning model is used in making prediction for all residues regardless of the categories of secondary structures they belong to. In this paper, we propose to train separate deep learning models for each category of secondary structures. Machine learning methods strive to achieve generality over the training examples and consequently loose accuracy. In this work, we explicitly exploit classification knowledge to restrict generalisation within the specific class of training examples. This is to compensate the loss of generalisation by exploiting specialisation knowledge in an informed way.

Results

The new method named SAP4SS obtains mean absolute error (MAE) values of 15.59, 18.87, 6.03, and 21.71 respectively for four types of backbone angles \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\phi$$\end{document}ϕ, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\psi$$\end{document}ψ, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\theta$$\end{document}θ, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\tau$$\end{document}τ. Consequently, SAP4SS significantly outperforms existing state-of-the-art methods SAP, OPUS-TASS, and SPOT-1D: the differences in MAE for all four types of angles are from 1.5 to 4.1% compared to the best known results.

Availability

SAP4SS along with its data is available from https://gitlab.com/mahnewton/sap4ss.

Information quantity for secondary structure propensities of protein subsequences in the Protein Data Bank

Elucidating the principles of sequence–structure relationships of proteins is a long-standing issue in biology. The nature of a short segment of a protein is determined by both the subsequence of the segment itself and its environment. For example, a type of subsequence, the so-called chameleon sequences, can form different secondary structures depending on its environments. Chameleon sequences are considered to have a weak tendency to form a specific structure. Although many chameleon sequences have been identified, they are only a small part of all possible subsequences in the proteome. The strength of the tendency to take a specific structure for each subsequence has not been fully quantified. In this study, we comprehensively analyzed subsequences consisting of four to nine amino acid residues, or N-gram (4≤N≤9), observed in non-redundant sequences in the Protein Data Bank (PDB). Tendencies to form a specific structure in terms of the secondary structure and accessible surface area are quantified as information quantities for each N-gram. Although the majority of observed subsequences have low information quantity due to lack of samples in the current PDB, thousands of N-grams with strong tendencies, including known structural motifs, were found. In addition, machine learning partially predicted the tendency of unknown N-grams, and thus, this technique helps to extract knowledge from the limited number of samples in the PDB.

Structural Basis for Developing Multitarget Compounds Acting on Cysteinyl Leukotriene Receptor 1 and G-Protein-Coupled Bile Acid Receptor 1

G-protein-coupled receptors (GPCRs) are the molecular target of 40% of marketed drugs and the most investigated structures to develop novel therapeutics. Different members of the GPCRs superfamily can modulate the same cellular process acting on diverse pathways, thus representing an attractive opportunity to achieve multitarget drugs with synergic pharmacological effects. Here, we present a series of compounds with dual activity toward cysteinyl leukotriene receptor 1 (CysLT₁R) and G-protein-coupled bile acid receptor 1 (GPBAR1). They are derivatives of REV5901─the first reported dual compound─with therapeutic potential in the treatment of colitis and other inflammatory processes. We report the binding mode of the most active compounds in the two GPCRs, revealing unprecedented structural basis for future drug design studies, including the presence of a polar group opportunely spaced from an aromatic ring in the ligand to interact with Arg79^2.60 of CysLT₁R and achieve dual activity.

TopProperty: Robust Metaprediction of Transmembrane and Globular Protein Features Using Deep Neural Networks

Transmembrane proteins (TMPs) are critical components of cellular life. However, due to experimental challenges, the number of experimentally resolved TMP structures is severely underrepresented in databases compared to their cellular abundance. Prediction of (per-residue) features such as transmembrane topology, membrane exposure, secondary structure, and solvent accessibility can be a useful starting point for experimental design or protein structure prediction but often requires different computational tools for different features or types of proteins. We present TopProperty, a metapredictor that predicts all of these features for TMPs or globular proteins. TopProperty is trained on datasets without bias toward a high number of sequence homologs, and the predictions are significantly better than the evaluated state-of-the-art primary predictors on all quality metrics. TopProperty eliminates the need for protein type- or feature-tailored tools, specifically for TMPs. TopProperty is freely available as a web server and standalone at https://cpclab.uni-duesseldorf.de/topsuite/.

Increasing the accuracy of single sequence prediction methods using a deep semi-supervised learning framework

MOTIVATION

Over the past 50 years, our ability to model protein sequences with evolutionary information has progressed in leaps and bounds. However, even with the latest deep learning methods, the modelling of a critically important class of proteins, single orphan sequences, remains unsolved.

RESULTS

By taking a bioinformatics approach to semi-supervised machine learning, we develop Profile Augmentation of Single Sequences (PASS), a simple but powerful framework for building accurate single-sequence methods. To demonstrate the effectiveness of PASS we apply it to the mature field of secondary structure prediction. In doing so we develop S4PRED, the successor to the open-source PSIPRED-Single method, which achieves an unprecedented Q3 score of 75.3% on the standard CB513 test. PASS provides a blueprint for the development of a new generation of predictive methods, advancing our ability to model individual protein sequences.

AVAILABILITY AND IMPLEMENTATION

The S4PRED model is available as open source software on the PSIPRED GitHub repository (https://github.com/psipred/s4pred), along with documentation. It will also be provided as a part of the PSIPRED web service (http://bioinf.cs.ucl.ac.uk/psipred/).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Discovering the Ultimate Limits of Protein Secondary Structure Prediction

Secondary structure prediction (SSP) of proteins is an important structural biology technique with many applications. There have been ~300 algorithms published in the past seven decades with fierce competition in accuracy. In the first 60 years, the accuracy of three-state SSP rose from ~56% to 81%; after that, it has long stayed at 81-86%. In the 1990s, the theoretical limit of three-state SSP accuracy had been estimated to be 88%. Thus, SSP is now generally considered not challenging or too challenging to improve. However, we found that the limit of three-state SSP might be underestimated. Besides, there is still much room for improving segment-based and eight-state SSPs, but the limits of these emerging topics have not been determined. This work performs large-scale sequence and structural analyses to estimate SSP accuracy limits and assess state-of-the-art SSP methods. The limit of three-state SSP is re-estimated to be ~92%, 4-5% higher than previously expected, indicating that SSP is still challenging. The estimated limit of eight-state SSP is 84-87%. Several proposals for improving future SSP algorithms are made based on our results. We hope that these findings will help move forward the development of SSP and all its applications.

Improved protein relative solvent accessibility prediction using deep multi-view feature learning framework

The accurate prediction of the relative solvent accessibility of a protein is critical to understanding its 3D structure and biological function. In this study, a novel deep multi-view feature learning (DMVFL) framework that integrates three different neural network units, i.e., bidirectional long short-term memory recurrent neural network, squeeze-and-excitation, and fully-connected hidden layer, with four sequence-based single-view features, i.e., position-specific scoring matrix, position-specific frequency matrix, predicted secondary structure, and roughly predicted three-state relative solvent accessibility probability, is developed to accurately predict relative solvent accessibility information of protein. On the basis of this newly developed framework, one new protein relative solvent accessibility predictor was proposed and called DMVFL-RSA, which employs a customized multiple feedback mechanism that helps to extract discriminative information embedded in the four single-view features. In benchmark tests on TEST524 and CASP14-derived (CASP14set) datasets, DMVFL-RSA outperforms other existing state-of-the-art protein relative solvent accessibility predictors when predicting two-state (exposure threshold of 25%), three-state (exposure thresholds of 9% and 36%), and four-state (exposure thresholds of 4%, 25%, and 50%) discrete values. For real-valued prediction on TEST524 and CASP14set, DMVFL-RSA has also gained high Pearson correlation coefficient values, indicating a positive correlation between the predicted and native relative solvent accessibility. Detailed analyses show that the major advantages of DMVFL-RSA lie in the high efficiency of the DMVFL framework, the applied multiple feedback mechanism, and the strong sensitivity of the sequence-based features. The web server of DMVFL-RSA is freely available at https://jun-csbio.github.io/DMVFL-RSA/for academic use. The standalone package of DMVFL-RSA is downloadable at https://github.com/XueQiangFan/DMVFL-RSA.

A high-throughput predictive method for sequence-similar fold switchers

Although most experimentally characterized proteins with similar sequences assume the same folds and perform similar functions, an increasing number of exceptions is emerging. One class of exceptions comprises sequence-similar fold switchers, whose secondary structures shift from α-helix <-> β-sheet through a small number of mutations, a sequence insertion, or a deletion. Predictive methods for identifying sequence-similar fold switchers are desirable because some are associated with disease and/or can perform different functions in cells. Here, we use homology-based secondary structure predictions to identify sequence-similar fold switchers from their amino acid sequences alone. To do this, we predicted the secondary structures of sequence-similar fold switchers using three different homology-based secondary structure predictors: PSIPRED, JPred4, and SPIDER3. We found that α-helix <-> β-strand prediction discrepancies from JPred4 discriminated between the different conformations of sequence-similar fold switchers with high statistical significance (P < 1.8*10-19 ). Thus, we used these discrepancies as a classifier and found that they can often robustly discriminate between sequence-similar fold switchers and sequence-similar proteins that maintain the same folds (Matthews Correlation Coefficient of 0.82). We found that JPred4 is a more robust predictor of sequence-similar fold switchers because of (a) the curated sequence database it uses to produce multiple sequence alignments and (b) its use of sequence profiles based on Hidden Markov Models. Our results indicate that inconsistencies between JPred4 secondary structure predictions can be used to identify some sequence-similar fold switchers from their sequences alone. Thus, the negative information from inconsistent secondary structure predictions can potentially be leveraged to identify sequence-similar fold switchers from the broad base of genomic sequences.

How do I get the most out of my protein sequence using bioinformatics tools?

Biochemical and biophysical experiments are essential for uncovering the three-dimensional structure and biological role of a protein of interest. However, meaningful predictions can frequently also be made using bioinformatics resources that transfer knowledge from a well studied protein to an uncharacterized protein based on their evolutionary relatedness. These predictions are helpful in developing specific hypotheses to guide wet-laboratory experiments. Commonly used bioinformatics resources include methods to identify and predict conserved sequence motifs, protein domains, transmembrane segments, signal sequences, and secondary as well as tertiary structure. Here, several such methods available through the MPI Bioinformatics Toolkit (https://toolkit.tuebingen.mpg.de) are described and how their combined use can provide meaningful information on a protein of unknown function is demonstrated. In particular, the identification of homologs of known structure using HHpred, internal repeats using HHrepID, coiled coils using PCOILS and DeepCoil, and transmembrane segments using Quick2D are focused on.

A secondary structure-based position-specific scoring matrix applied to the improvement in protein secondary structure prediction

Protein secondary structure prediction (SSP) has a variety of applications; however, there has been relatively limited improvement in accuracy for years. With a vision of moving forward all related fields, we aimed to make a fundamental advance in SSP. There have been many admirable efforts made to improve the machine learning algorithm for SSP. This work thus took a step back by manipulating the input features. A secondary structure element-based position-specific scoring matrix (SSE-PSSM) is proposed, based on which a new set of machine learning features can be established. The feasibility of this new PSSM was evaluated by rigid independent tests with training and testing datasets sharing <25% sequence identities. In all experiments, the proposed PSSM outperformed the traditional amino acid PSSM. This new PSSM can be easily combined with the amino acid PSSM, and the improvement in accuracy was remarkable. Preliminary tests made by combining the SSE-PSSM and well-known SSP methods showed 2.0% and 5.2% average improvements in three- and eight-state SSP accuracies, respectively. If this PSSM can be integrated into state-of-the-art SSP methods, the overall accuracy of SSP may break the current restriction and eventually bring benefit to all research and applications where secondary structure prediction plays a vital role during development. To facilitate the application and integration of the SSE-PSSM with modern SSP methods, we have established a web server and standalone programs for generating SSE-PSSM available at http://10.life.nctu.edu.tw/SSE-PSSM.

Predicting Protein Therapeutic Candidates for Bovine Babesiosis Using Secondary Structure Properties and Machine Learning

Bovine babesiosis causes significant annual global economic loss in the beef and dairy cattle industry. It is a disease instigated from infection of red blood cells by haemoprotozoan parasites of the genus Babesia in the phylum Apicomplexa. Principal species are Babesia bovis, Babesia bigemina, and Babesia divergens. There is no subunit vaccine. Potential therapeutic targets against babesiosis include members of the exportome. This study investigates the novel use of protein secondary structure characteristics and machine learning algorithms to predict exportome membership probabilities. The premise of the approach is to detect characteristic differences that can help classify one protein type from another. Structural properties such as a protein’s local conformational classification states, backbone torsion angles ϕ (phi) and ψ (psi), solvent-accessible surface area, contact number, and half-sphere exposure are explored here as potential distinguishing protein characteristics. The presented methods that exploit these structural properties via machine learning are shown to have the capacity to detect exportome from non-exportome Babesia bovis proteins with an 86–92% accuracy (based on 10-fold cross validation and independent testing). These methods are encapsulated in freely available Linux pipelines setup for automated, high-throughput processing. Furthermore, proposed therapeutic candidates for laboratory investigation are provided for B. bovis, B. bigemina, and two other haemoprotozoan species, Babesia canis, and Plasmodium falciparum.

The influence of dataset homology and a rigorous evaluation strategy on protein secondary structure prediction

The secondary structure prediction (SSP) of proteins has long been an essential structural biology technique with various applications. Despite its vital role in many research and industrial fields, in recent years, as the accuracy of state-of-the-art secondary structure predictors approaches the theoretical upper limit, SSP has been considered no longer challenging or too challenging to make advances. With the belief that the substantial improvement of SSP will move forward many fields depending on it, we conducted this study, which focused on three issues that have not been noticed or thoroughly examined yet but may have affected the reliability of the evaluation of previous SSP algorithms. These issues are all about the sequence homology between or within the developmental and evaluation datasets. We thus designed many different homology layouts of datasets to train and evaluate SSP prediction models. Multiple repeats were performed in each experiment by random sampling. The conclusions obtained with small experimental datasets were verified with large-scale datasets using state-of-the-art SSP algorithms. Very different from the long-established assumption, we discover that the sequence homology between query datasets for training, testing, and independent tests exerts little influence on SSP accuracy. Besides, the sequence homology redundancy between or within most datasets would make the accuracy of an SSP algorithm overestimated, while the redundancy within the reference dataset for extracting predictive features would make the accuracy underestimated. Since the overestimating effects are more significant than the underestimating effect, the accuracy of some SSP methods might have been overestimated. Based on the discoveries, we propose a rigorous procedure for developing SSP algorithms and making reliable evaluations, hoping to bring substantial improvements to future SSP methods and benefit all research and application fields relying on accurate prediction of protein secondary structures.

Machine learning designs non-hemolytic antimicrobial peptides

Machine learning (ML) consists of the recognition of patterns from training data and offers the opportunity to exploit large structure-activity databases for drug design. In the area of peptide drugs, ML is mostly being tested to design antimicrobial peptides (AMPs), a class of biomolecules potentially useful to fight multidrug-resistant bacteria. ML models have successfully identified membrane disruptive amphiphilic AMPs, however mostly without addressing the associated toxicity to human red blood cells. Here we trained recurrent neural networks (RNN) with data from DBAASP (Database of Antimicrobial Activity and Structure of Peptides) to design short non-hemolytic AMPs. Synthesis and testing of 28 generated peptides, each at least 5 mutations away from training data, allowed us to identify eight new non-hemolytic AMPs against Pseudomonas aeruginosa, Acinetobacter baumannii, and methicillin-resistant Staphylococcus aureus (MRSA). These results show that machine learning (ML) can be used to design new non-hemolytic AMPs.

Variable Length Character N-Gram Embedding of Protein Sequences for Secondary Structure Prediction

BACKGROUND

The prediction of a protein's secondary structure from its amino acid sequence is an essential step towards predicting its 3-D structure. The prediction performance improves by incorporating homologous multiple sequence alignment information. Since homologous details not available for all proteins. Therefore, it is necessary to predict the protein secondary structure from single sequences.

OBJECTIVE AND METHODS

Protein secondary structure predicted from their primary sequences using n-gram word embedding and deep recurrent neural network. Protein secondary structure depends on local and long-range neighbor residues in primary sequences. In the proposed work, the local contextual information of amino acid residues captures variable-length character n-gram words. An embedding vector represents these variable-length character n-gram words. Further, the bidirectional long short-term memory (Bi-LSTM) model is used to capture the long-range contexts by extracting the past and future residues information in primary sequences.

RESULTS

The proposed model evaluates on three public datasets ss.txt, RS126, and CASP9. The model shows the Q3 accuracy of 92.57%, 86.48%, and 89.66% for ss.txt, RS126, and CASP9.

CONCLUSION

The proposed model performance compares with state-of-the-art methods available in the literature. After a comparative analysis, it observed that the proposed model performs better than state-of-the-art methods.

Emergence of Hybrid Resistance and Virulence Plasmids Harboring New Delhi Metallo-β-Lactamase in Klebsiella pneumoniae in Russia

The emergence of carbapenem-resistant hypervirulent Klebsiella pneumoniae (CR-hvKp) is a new threat to healthcare. In this study, we analyzed nine CR-hvKp isolates of different sequence-types (ST) recovered from patients with nosocomial infections in two hospitals in Saint Petersburg. Whole-genome sequencing showed that eight of them harbored large mosaic plasmids carrying resistance to carbapenems and hypervirulence simultaneously, and four different types of hybrid plasmids were identified. BLAST analysis showed a high identity with two hybrid plasmids originating in the UK and Czech Republic. We demonstrated that hybrid plasmids emerged due to the acquisition of resistance genes by virulent plasmids. Moreover, one of the hybrid plasmids carried a novel New Delhi metallo-beta-lactamase (NDM) variant, differing from NDM-1 by one amino acid substitution (D130N), which did not provide significant evolutionary advantages compared to NDM-1. The discovery of structurally similar plasmids in geographically distant regions suggests that the actual distribution of hybrid plasmids carrying virulence and resistance genes is much wider than expected.

A novel de novo NIPA1 missense mutation associated to hereditary spastic paraplegia

SPG6 accounts for 1% of autosomal dominant Hereditary Spastic Paraplegia (HSP) and is caused by pathogenic variants in NIPA1, which encodes a magnesium transporter located in plasma membrane and early endosomes, implicated in neuronal development and maintenance. Here we report a 39-year-old woman affected by progressive gait disturbance associated to absence seizures episodes within childhood. Clinical exome sequencing identified a likely pathogenic de novo heterozygous variant in NIPA1 (NM_144599.5 c.249 C > G; p.Asn83Lys). Molecular modelling was performed to evaluate putative functional consequence of the NIPA1 protein. Indeed, the Asn83Lys modification is predicted to induce a significant perturbation of the protein structure, altering signal transduction or small-molecule transport by modulating the length of the second transmembrane domain. This is the first study reporting a SPG6-affected patient harbouring the NIPA1 p.Asn83Lys mutation.

Recent Applications of Deep Learning Methods on Evolution- and Contact-Based Protein Structure Prediction

The new advances in deep learning methods have influenced many aspects of scientific research, including the study of the protein system. The prediction of proteins' 3D structural components is now heavily dependent on machine learning techniques that interpret how protein sequences and their homology govern the inter-residue contacts and structural organization. Especially, methods employing deep neural networks have had a significant impact on recent CASP13 and CASP14 competition. Here, we explore the recent applications of deep learning methods in the protein structure prediction area. We also look at the potential opportunities for deep learning methods to identify unknown protein structures and functions to be discovered and help guide drug-target interactions. Although significant problems still need to be addressed, we expect these techniques in the near future to play crucial roles in protein structural bioinformatics as well as in drug discovery.

RNA Backbone Torsion and Pseudotorsion Angle Prediction Using Dilated Convolutional Neural Networks

RNA three-dimensional structure prediction has been relied on using a predicted or experimentally determined secondary structure as a restraint to reduce the conformational sampling space. However, the secondary-structure restraints are limited to paired bases, and the conformational space of the ribose-phosphate backbone is still too large to be sampled efficiently. Here, we employed the dilated convolutional neural network to predict backbone torsion and pseudotorsion angles using a single RNA sequence as input. The method called SPOT-RNA-1D was trained on a high-resolution training data set and tested on three independent, nonredundant, and high-resolution test sets. The proposed method yields substantially smaller mean absolute errors than the baseline predictors based on random predictions and based on helix conformations according to actual angle distributions. The mean absolute errors for three test sets range from 14°-44° for different angles, compared to 17°-62° by random prediction and 14°-58° by helix prediction. The method also accurately recovers the overall patterns of single or pairwise angle distributions. In general, torsion angles further away from the bases and associated with unpaired bases and paired bases involved in tertiary interactions are more difficult to predict. Compared to the best models in RNA-puzzles experiments, SPOT-RNA-1D yielded more accurate dihedral angles and, thus, are potentially useful as model quality indicators and restraints for RNA structure prediction as in protein structure prediction.

SPOT-1D-Single: Improving the Single-Sequence-Based Prediction of Protein Secondary Structure, Backbone Angles, Solvent Accessibility and Half-Sphere Exposures using a Large Training Set and Ensembled Deep Learning

MOTIVATION

Knowing protein secondary and other one-dimensional structural properties are essential for accurate protein structure and function prediction. As a result, many methods have been developed for predicting these one-dimensional structural properties. However, most methods relied on evolutionary information that may not exist for many proteins due to a lack of sequence homologs. Moreover, it is computationally intensive for obtaining evolutionary information as the library of protein sequences continues to expand exponentially. Here we developed a new single-sequence method called SPOT-1D-Single based on a large training dataset of 39120 proteins deposited prior to 2016 and an ensemble of hybrid Long-Short-Term-Memory bidirectional neural network and convolutional neural network.

RESULTS

We showed that SPOT-1D-Single consistently improves over SPIDER3-Single and ProteinUnet for secondary structure, solvent accessibility, contact number, and backbone angles prediction for all seven independent test sets (TEST2018, SPOT-2016, SPOT-2016-HQ, SPOT-2018, SPOT-2018-HQ, CASP12, and CASP13 free-modeling targets). For example, the predicted three-state secondary structure's accuracy ranges from 72.12-74.28% by SPOT-1D-Single, compared to 69.1-72.6% by SPIDER3-Single and 70.6-73% by ProteinUnet. SPOT-1D-Single also predicts SS3 and SS8 with 6.24% and 6.98% better accuracy than SPOT-1D on SPOT-2018 proteins with no homologs (Neff=1), respectively. The new method's improvement over existing techniques is due to a larger training set combined with ensembled learning.

AVAILABILITY

Standalone-version of SPOT-1D-Single is available at https://github.com/jas-preet/SPOT-1D-Single. Direct prediction can also be made at https://sparks-lab.org/server/spot-1d-single. The datasets used in this research can also be downloaded from GitHub.

Protein Secondary Structure Prediction Using Character Bi-gram Embedding and Bi-LSTM

Background:

Protein secondary structure is vital to predicting the tertiary structure, which is essential in deciding protein function and drug designing. Therefore, there is a high requirement of computational methods to predict secondary structure from their primary sequence. Protein primary sequences represented as a linear combination of twenty amino acid characters and contain the contextual information for secondary structure prediction.

Objective and Methods:

Protein secondary structure predicted from their primary sequences using a deep recurrent neural network. Protein secondary structure depends on local and long-range residues in primary sequences. In the proposed work, the local contextual information of amino acid residues captures with character n-gram. A dense embedding vector represents this local contextual information. Furthermore, the bidirectional long short-term memory (Bi-LSTM) model is used to capture the long-range contexts by extracting the past and future residues information in primary sequences.

Results:

The proposed deep recurrent architecture is evaluated for its efficacy for datasets, namely ss.txt, RS126, and CASP9. The model shows the Q3 accuracies of 88.45%, 83.48%, and 86.69% for ss.txt, RS126, and CASP9, respectively. The performance of the proposed model is also compared with other state-of-the-art methods available in the literature.

Conclusion:

After a comparative analysis, it was observed that the proposed model is performing better in comparison to state-of-art methods.

Multiobjective heuristic algorithm for de novo protein design in a quantified continuous sequence space

Protein design usually involves sequence search process and evaluation criteria. Commonly used methods primarily implement the Monte Carlo or simulated annealing algorithm with a single-energy function to obtain ideal solutions, which is often highly time-consuming and limited by the accuracy of the energy function. In this report, we introduce a multiobjective algorithm named Hydra for protein design, which employs two different energy functions to optimize solutions simultaneously and makes use of the latent quantitative relationship between different amino acid types to facilitate the search process. The framework uses two kinds of prior information to transform the original disordered discrete sequence space into a relatively ordered space, and decoy sequences are searched in this ordered space through a multiobjective swarm intelligence algorithm. This algorithm features high accuracy and a high-speed search process. Our method was tested on 40 targets covering different fold classes, which were computationally verified to be well folded, and it experimentally solved the 1UBQ fold by NMR in excellent agreement with the native structure with a backbone RMSD deviation of 1.074 Å. The Hydra software package can be downloaded from: http://www.csbio.sjtu.edu.cn/bioinf/HYDRA/ for academic use.

Prediction and analysis of multiple protein lysine modified sites based on conditional wasserstein generative adversarial networks

BACKGROUND

Protein post-translational modification (PTM) is a key issue to investigate the mechanism of protein's function. With the rapid development of proteomics technology, a large amount of protein sequence data has been generated, which highlights the importance of the in-depth study and analysis of PTMs in proteins.

METHOD

We proposed a new multi-classification machine learning pipeline MultiLyGAN to identity seven types of lysine modified sites. Using eight different sequential and five structural construction methods, 1497 valid features were remained after the filtering by Pearson correlation coefficient. To solve the data imbalance problem, Conditional Generative Adversarial Network (CGAN) and Conditional Wasserstein Generative Adversarial Network (CWGAN), two influential deep generative methods were leveraged and compared to generate new samples for the types with fewer samples. Finally, random forest algorithm was utilized to predict seven categories.

RESULTS

In the tenfold cross-validation, accuracy (Acc) and Matthews correlation coefficient (MCC) were 0.8589 and 0.8376, respectively. In the independent test, Acc and MCC were 0.8549 and 0.8330, respectively. The results indicated that CWGAN better solved the existing data imbalance and stabilized the training error. Alternatively, an accumulated feature importance analysis reported that CKSAAP, PWM and structural features were the three most important feature-encoding schemes. MultiLyGAN can be found at https://github.com/Lab-Xu/MultiLyGAN .

CONCLUSIONS

The CWGAN greatly improved the predictive performance in all experiments. Features derived from CKSAAP, PWM and structure schemes are the most informative and had the greatest contribution to the prediction of PTM.

Enhancing protein backbone angle prediction by using simpler models of deep neural networks

Protein structure prediction is a grand challenge. Prediction of protein structures via the representations using backbone dihedral angles has recently achieved significant progress along with the on-going surge of deep neural network (DNN) research in general. However, we observe that in the protein backbone angle prediction research, there is an overall trend to employ more and more complex neural networks and then to throw more and more features to the neural networks. While more features might add more predictive power to the neural network, we argue that redundant features could rather clutter the scenario and more complex neural networks then just could counterbalance the noise. From artificial intelligence and machine learning perspectives, problem representations and solution approaches do mutually interact and thus affect performance. We also argue that comparatively simpler predictors can more easily be reconstructed than the more complex ones. With these arguments in mind, we present a deep learning method named Simpler Angle Predictor (SAP) to train simpler DNN models that enhance protein backbone angle prediction. We then empirically show that SAP significantly outperforms existing state-of-the-art methods on well-known benchmark datasets: for some types of angles, the differences are above 3 in mean absolute error (MAE). The SAP program along with its data is available from the website https://gitlab.com/mahnewton/sap.

ProteinUnet-An efficient alternative to SPIDER3-single for sequence-based prediction of protein secondary structures

Predicting protein function and structure from sequence remains an unsolved problem in bioinformatics. The best performing methods rely heavily on evolutionary information from multiple sequence alignments, which means their accuracy deteriorates for sequences with a few homologs, and given the increasing sequence database sizes requires long computation times. Here, a single‐sequence‐based prediction method is presented, called ProteinUnet, leveraging an U‐Net convolutional network architecture. It is compared to SPIDER3‐Single model, based on long short‐term memory‐bidirectional recurrent neural networks architecture. Both methods achieve similar results for prediction of secondary structures (both three‐ and eight‐state), half‐sphere exposure, and contact number, but ProteinUnet has two times fewer parameters, 17 times shorter inference time, and can be trained 11 times faster. Moreover, ProteinUnet tends to be better for short sequences and residues with a low number of local contacts. Additionally, the method of loss weighting is presented as an effective way of increasing accuracy for rare secondary structures.