Search

Therapeutic antibodies are one of the most important parts of the pharmaceutical industry. They are widely used in treating various diseases such as autoimmune diseases, cancer, inflammation, and infectious diseases. Their development process however is often brought to a standstill or takes a longer time and is then more expensive due to their hydrophobicity problems. Hydrophobic interactions can cause problems on half-life, drug administration, and immunogenicity at all stages of antibody drug development. Some of the most widely accepted and used technologies for determining the hydrophobic interactions of antibodies include standup monolayer adsorption chromatography (SMAC), salt-gradient affinity-capture self-interaction nanoparticle spectroscopy (SGAC-SINS), and hydrophobic interaction chromatography (HIC). However, to measure SMAC, SGAC-SINS, and HIC for hundreds of antibody drug candidates is time-consuming and costly. To save time and money, a predictor called SSH is developed. Based on the antibody’s sequence only, it can predict the hydrophobic interactions of monoclonal antibodies (mAbs). Using the leave-one-out crossvalidation, SSH achieved 91.226% accuracy, 96.396% sensitivity or recall, 84.196% specificity, 87.754% precision, 0.828 Mathew correlation coefficient (MCC), 0.919 f-score, and 0.961 area under the receiver operating characteristic (ROC) curve (AUC).

The field of proteomics has expanded recently with more sensitive techniques for the bulk measurement of peptides as well as single-molecule techniques. One limiting factor for some of these methods is the need for multiple chemical derivatizations and highly pure proteins free of contaminants. We demonstrate a solid-phase capture strategy suitable for the proteolysis, purification, and subsequent chemical modification of peptides. We use this resin on an HEK293T cell lysate and perform one-pot proteolysis, capture, and derivatization to generate a cellular proteome that identified over 40,000 bead-bound peptides. We also show that this capture can be reversed in a traceless manner, such that it is amenable for single-molecule proteomics techniques. With this technique, we perform a fluorescent labeling and C-terminal derivatization on a peptide and subject it to fluorosequencing, demonstrating that washing the resin is sufficient to remove excess dyes and other reagents prior to single-molecule protein sequencing.

Zebra2 is a highly automated web-tool to search for subfamily-specific and conserved positions (i.e. the determinants of functional diversity as well as the key catalytic and structural residues) in protein superfamilies. The bioinformatic analysis is facilitated by Mustguseal—a companion web-server to automatically collect and superimpose a large representative set of functionally diverse homologs with high structure similarity but low sequence identity to the selected query protein. The results are automatically prioritized and provided at four information levels to facilitate the knowledge-driven expert selection of the most promising positions on-line: as a sequence similarity network; interfaces to sequence-based and 3D-structure-based analysis of conservation and variability; and accompanied by the detailed annotation of proteins accumulated from the integrated databases with links to the external resources. The integration of Zebra2 and Mustguseal web-tools provides the first of its kind out-of-the-box open-access solution to conduct a systematic analysis of evolutionarily related proteins implementing different functions within a shared 3D-structure of the superfamily, determine common and specific patterns of function-associated local structural elements, assist to select hot-spots for rational design and to prepare focused libraries for directed evolution. The web-servers are free and open to all users at https://biokinet.belozersky.msu.ru/zebra2, no login required.

In vitro selection using mRNA display is currently a widely used method to isolate functional peptides with desired properties. The analysis of High Throughput Sequencing (HTS) data from in vitro evolution experiments has proven to be a powerful technique but only recently has it been applied to mRNA display selections. In this Perspective, we introduce aspects of mRNA display and HTS that may be of interest to physical chemists. We highlight the potential of HTS to analyze in vitro selections of peptides and review recent advances in the application of HTS analysis to mRNA display experiments. We discuss some possible issues involved with HTS analysis and summarize some strategies to alleviate them. Finally, the potential for future impact of advancing HTS analysis on mRNA display experiments is discussed., Graphical Abstract High-throughput sequencing (HTS) of mRNA display selection of functional peptides

Automated Edman degradation for sequencing is the classic tool for determining the N-terminal sequences of proteins. Although it has been complemented or largely replaced by a variety of mass spectrometric methods in recent years, there is no simple mass spectrometric method for identification and de novo determination of the N-terminal sequence of a protein. Here, we present a simple and reliable approach which identifies N-terminal peptides of proteins by selectively labeling the N-terminal α-amino groups with a fluorescent reagent followed by enzyme digestion and tandem mass spectrometric analysis using an LC-MS/MS system coupled with a fluorescence detector. This approach also includes a differential peptide mapping step for identification of pyroglutamate-blocked N-terminal peptides. In addition, the strategy can be applied to proteins in gel. With this approach, complete N-terminal sequences of proteins can be obtained precisely and efficiently.

International audience; Rigorous validation of amino acid sequence is fundamental in the characterization of original and biosimilar protein biopharmaceuticals. Widely accepted workflows are based on bottom-up mass spectrometry, and they often require multiple techniques and significant manual work. Here, we demonstrate that optimization of a set of tandem mass spectroscopy (MS/MS) collision energies and automated combination of all available information in the measurements can increase the sequence validated by one technique close to the inherent limits. We created a software (called “Serac”) that consumes results of the Mascot database search engine and identifies the amino acids validated by bottom-up MS/MS experiments using the most rigorous, industrially acceptable definition of sequence coverage (we term this “confirmed sequence coverage”). The software can combine spectra at the level of amino acids or fragment ions to exploit complementarity, provides full transparency to justify validation, and reduces manual effort. With its help, we investigated collision energy dependence of confirmed sequence coverage of individual peptides and full proteins on trypsin-digested monoclonal antibody samples (rituximab and trastuzumab). We found the energy dependence to be modest, but we demonstrated the benefit of using spectra taken at multiple energies. We describe a workflow based on 2–3 LC-MS/MS runs, carefully selected collision energies, and a fragment ion level combination, which yields ∼85% confirmed sequence coverage, 25%–30% above that from a basic proteomics protocol. Further increase can mainly be expected from alternative digestion enzymes or fragmentation techniques, which can be seamlessly integrated to the processing, thereby allowing effortless validation of full sequences.

Protein-polymer bioconjugate self-assembly has attracted a great deal of attention as a method to fabricate protein nanomaterials in solution and the solid state. To identify protein properties that affect phase behavior in protein-polymer block copolymers, a library of 15 unique protein-b-poly(N-isopropylacrylamide) (PNIPAM) copolymers comprising 11 different proteins was compiled and analyzed. Many attributes of phase behavior are found to be similar among all studied bioconjugates regardless of protein properties, such as formation of micellar phases at high temperature and low concentration, lamellar ordering with increasing temperature, and disordering at high concentration, but several key protein-dependent trends are also observed. In particular, hexagonal phases are only observed for proteins within the molar mass range 20-36 kDa, where ordering quality is also significantly enhanced. While ordering is generally found to improve with increasing molecular weight outside of this range, most large bioconjugates exhibited weaker than predicted assembly, which is attributed to chain entanglement with increasing polymer molecular weight. Additionally, order-disorder transition boundaries are found to be largely uncorrelated to protein size and quality of ordering. However, the primary finding is that bioconjugate ordering can be accurately predicted using only protein molecular weight and percentage of residues contained within β sheets. This model provides a basis for designing protein-PNIPAM bioconjugates that exhibit well-defined self-assembly and a modeling framework that can generalize to other bioconjugate chemistries., United States. Department of Energy. Office of Basic Energy Sciences (Award DE-SC0007106)

Motivation Almost all protein residue contact prediction methods rely on the availability of deep multiple sequence alignments (MSAs). However, many proteins from the poorly populated families do not have sufficient number of homologs in the conventional UniProt database. Here we aim to solve this issue by exploring the rich sequence data from the metagenome sequencing projects. Results Based on the improved MSA constructed from the metagenome sequence data, we developed MapPred, a new deep learning-based contact prediction method. MapPred consists of two component methods, DeepMSA and DeepMeta, both trained with the residual neural networks. DeepMSA was inspired by the recent method DeepCov, which was trained on 441 matrices of covariance features. By considering the symmetry of contact map, we reduced the number of matrices to 231, which makes the training more efficient in DeepMSA. Experiments show that DeepMSA outperforms DeepCov by 10–13% in precision. DeepMeta works by combining predicted contacts and other sequence profile features. Experiments on three benchmark datasets suggest that the contribution from the metagenome sequence data is significant with P-values less than 4.04E-17. MapPred is shown to be complementary and comparable the state-of-the-art methods. The success of MapPred is attributed to three factors: the deeper MSA from the metagenome sequence data, improved feature design in DeepMSA and optimized training by the residual neural networks. Availability and implementation http://yanglab.nankai.edu.cn/mappred/. Supplementary information Supplementary data are available at Bioinformatics online.

The selection of proteotypic peptides, that is, detectable unique representatives of proteins of interest, is a key step in targeted proteomics. To date, much effort has been made to understand the mechanisms underlying peptide detection in liquid chromatography-tandem mass spectrometry (LC-MS/MS) based shotgun proteomics and to predict proteotypic peptides in the absence of experimental LC-MS/MS data. However, the prediction accuracy of existing tools is still unsatisfactory. We find that one crucial reason is their neglect of the significant influence of protein proteolytic digestion on peptide detectability in shotgun proteomics. Here, we present an Advanced Proteotypic Peptide Predictor (AP3), which explicitly takes peptide digestibility into account for the prediction of proteotypic peptides. Specifically, peptide digestibility is first predicted for each peptide and then incorporated as a feature into the peptide detectability prediction model. Our results demonstrated that peptide digestibility is the most important feature for the accurate prediction of proteotypic peptides in our model. Compared with the existing available algorithms, AP3 showed 10.3-34.7% higher prediction accuracy. On a targeted proteomics data set, AP3 accurately predicted the proteotypic peptides for proteins of interest, showing great potential for assisting the design of targeted proteomics experiments.

While protein sequence data is an emerging application domain for machine learning methods, small modifications to protein sequences can result in difficult-to-predict changes to the protein’s function. Consequently, protein machine learning models typically do not use randomized data augmentation procedures analogous to those used in computer vision or natural language, e.g., cropping or synonym substitution. In this paper, we empirically explore a set of simple string manipulations, which we use to augment protein sequence data when fine-tuning semi-supervised protein models. We provide 276 different comparisons to the Tasks Assessing Protein Embeddings (TAPE) baseline models, with Transformer-based models and training datasets that vary from the baseline methods only in the data augmentations and representation learning procedure. For each TAPE validation task, we demonstrate improvements to the baseline scores when the learned protein representation is fixed between tasks. We also show that contrastive learning fine-tuning methods typically outperform masked-token prediction in these models, with increasing amounts of data augmentation generally improving performance for contrastive learning protein methods. We find the most consistent results across TAPE tasks when using domain-motivated transformations, such as amino acid replacement, as well as restricting the Transformer attention to randomly sampled sub-regions of the protein sequence. In rarer cases, we even find that information-destroying augmentations, such as randomly shuffling entire protein sequences, can improve downstream performance.

CATH (https://www.cathdb.info) identifies domains in protein structures from wwPDB and classifies these into evolutionary superfamilies, thereby providing structural and functional annotations. There are two levels: CATH-B, a daily snapshot of the latest domain structures and superfamily assignments, and CATH+, with additional derived data, such as predicted sequence domains, and functionally coherent sequence subsets (Functional Families or FunFams). The latest CATH+ release, version 4.3, significantly increases coverage of structural and sequence data, with an addition of 65,351 fully-classified domains structures (+15%), providing 500 238 structural domains, and 151 million predicted sequence domains (+59%) assigned to 5481 superfamilies. The FunFam generation pipeline has been re-engineered to cope with the increased influx of data. Three times more sequences are captured in FunFams, with a concomitant increase in functional purity, information content and structural coverage. FunFam expansion increases the structural annotations provided for experimental GO terms (+59%). We also present CATH-FunVar web-pages displaying variations in protein sequences and their proximity to known or predicted functional sites. We present two case studies (1) putative cancer drivers and (2) SARS-CoV-2 proteins. Finally, we have improved links to and from CATH including SCOP, InterPro, Aquaria and 2DProt.

The topology of protein folds can be specified by the inter-residue contact-maps and accurate contact-map prediction can help ab initio structure folding. We developed TripletRes to deduce protein contact-maps from discretized distance profiles by end-to-end training of deep residual neural-networks. Compared to previous approaches, the major advantage of TripletRes is in its ability to learn and directly fuse a triplet of coevolutionary matrices extracted from the whole-genome and metagenome databases and therefore minimize the information loss during the course of contact model training. TripletRes was tested on a large set of 245 non-homologous proteins from CASP 11&12 and CAMEO experiments and outperformed other top methods from CASP12 by at least 58.4% for the CASP 11&12 targets and 44.4% for the CAMEO targets in the top-L long-range contact precision. On the 31 FM targets from the latest CASP13 challenge, TripletRes achieved the highest precision (71.6%) for the top-L/5 long-range contact predictions. It was also shown that a simple re-training of the TripletRes model with more proteins can lead to further improvement with precisions comparable to state-of-the-art methods developed after CASP13. These results demonstrate a novel efficient approach to extend the power of deep convolutional networks for high-accuracy medium- and long-range protein contact-map predictions starting from primary sequences, which are critical for constructing 3D structure of proteins that lack homologous templates in the PDB library., Author summary Ab initio protein folding has been a major unsolved problem in computational biology for more than half a century. Recent community-wide Critical Assessment of Structure Prediction (CASP) experiments have witnessed exciting progress on ab initio structure prediction, which was mainly powered by the boosting of contact-map prediction as the latter can be used as constraints to guide ab initio folding simulations. In this work, we proposed a new open-source deep-learning architecture, TripletRes, built on the residual convolutional neural networks for high-accuracy contact prediction. The large-scale benchmark and blind test results demonstrate competitive performance of the proposed methods to other top approaches in predicting medium- and long-range contact-maps that are critical for guiding protein folding simulations. Detailed data analyses showed that the major advantage of TripletRes lies in the unique protocol to fuse multiple evolutionary feature matrices which are directly extracted from whole-genome and metagenome databases and therefore minimize the information loss during the contact model training.

Lasso peptides are a diverse class of ribosomally synthesized and post-translationally modified peptides (RiPPs). Their proteolytic and thermal stability alongside their growing potential as therapeutics has increased attention to these antimicrobial peptides. With the advent of genome mining, the discovery of RiPPs allows for the accurate prediction of putatively encoded structures, however, MSn experiments only provide partial sequence confirmation, therefore 2D NMR experiments are necessary for characterisation. Multiple MS/MS techniques were applied to two structurally characterized lasso peptides, huascopeptin and leepeptin, and one uncharacterized lasso peptide, citrulassin C, which was not isolable in sufficient quantity for NMR analysis. We have shown that MS2 can be used to elucidate the full amino acid sequences previously predicted with genome mining for this compound class. HCD was able to open the macrocycles and fragment the newly opened linear peptides, confirming the complete amino acid sequences of the characterised lasso peptides. In addition, to determine if this technique could be applied at the earliest stages of the isolation process, we targeted a lasso peptide found by genome mining, citrulassin C, and were able to fully elucidate the amino acid sequence using only MS2 from a semi-crude extract of Streptomyces huasconensis HST28T.

The CRISPR-Cas are adaptive bacterial and archaeal immunity systems that have been harnessed for the development of powerful genome editing and engineering tools. In the incessant host-parasite arms race, viruses evolved multiple anti-defense mechanisms including diverse anti-CRISPR proteins (Acrs) that specifically inhibit CRISPR-Cas and therefore have enormous potential for application as modulators of genome editing tools. Most Acrs are small and highly variable proteins which makes their bioinformatic prediction a formidable task. We present a machine-learning approach for comprehensive Acr prediction. The model shows high predictive power when tested against an unseen test set and was employed to predict 2,500 candidate Acr families. Experimental validation of top candidates revealed two unknown Acrs (AcrIC9, IC10) and three other top candidates were coincidentally identified and found to possess anti-CRISPR activity. These results substantially expand the repertoire of predicted Acrs and provide a resource for experimental Acr discovery., CRISPR-Cas is a host adaptive immunity system and viruses harbor diverse anti-CRISPR proteins (Acrs). Here, the authors develop a random forest machine-learning approach to predict Acrs, identifying 2500 candidate Acr families, which expand the current repertoire of predicted Acrs by two orders of magnitude.

Antibodies are capable of potently and specifically binding individual antigens and, in some cases, disrupting their functions. The key challenge in generating antibody-based inhibitors is the lack of fundamental information relating sequences of antibodies to their unique properties as inhibitors. We develop a pipeline, Antibody Sequence Analysis Pipeline using Statistical testing and Machine Learning (ASAP-SML), to identify features that distinguish one set of antibody sequences from antibody sequences in a reference set. The pipeline extracts feature fingerprints from sequences. The fingerprints represent germline, CDR canonical structure, isoelectric point and frequent positional motifs. Machine learning and statistical significance testing techniques are applied to antibody sequences and extracted feature fingerprints to identify distinguishing feature values and combinations thereof. To demonstrate how it works, we applied the pipeline on sets of antibody sequences known to bind or inhibit the activities of matrix metalloproteinases (MMPs), a family of zinc-dependent enzymes that promote cancer progression and undesired inflammation under pathological conditions, against reference datasets that do not bind or inhibit MMPs. ASAP-SML identifies features and combinations of feature values found in the MMP-targeting sets that are distinct from those in the reference sets., Author summary The availability of machine learning techniques and the exponential growth of sequencing data presents new opportunities to identify features that endow antibodies with the ability to disrupt the functions of biological targets. We have created a pipeline that uses statistical testing and machine learning techniques to determine features that are overrepresented in a specified set of antibody sequences in comparison to a reference set. The pipeline is referred to as Antibody Sequence Analysis Pipeline using Statistical testing and Machine Learning (ASAP-SML). We demonstrate the use of ASAP-SML by analyzing sets of antibodies that inhibit matrix metalloproteinases (MMPs) against reference sets. ASAP-SML performs within and across set similarity analysis. As in prior studies, our analysis of these datasets shows that features associated with the antibody heavy chain are more likely to differentiate MMP-targeting antibody sequences from reference antibody sequences. Further, ASAP-SML identifies several features in the MMP-targeting set that are distinct from the reference sets. Using design recommendation trees, ASAP-SML suggests combinations of features that can be included or excluded to augment the targeting set with additional candidate MMP-targeting antibody sequences.

We present ProteoClade, a Python toolkit that performs taxa-specific peptide assignment, protein inference, and quantitation for multi-species proteomics experiments. ProteoClade scales to hundreds of millions of protein sequences, requires minimal computational resources, and is open source, multi-platform, and accessible to non-programmers. We demonstrate its utility for processing quantitative proteomic data derived from patient-derived xenografts and its speed and scalability enable a novel de novo proteomic workflow for complex microbiota samples., Author summary The exponential growth of the number of available reference protein sequences has provided an opportunity to taxonomically annotate and quantify complex mixtures of organisms using bottom-up proteomics. However, the ability to annotate relevant taxa to proteomics data is computationally challenging when data sets generate millions of candidate sequences and the reference database contains billions of peptide sequences. Here, we provide a software tool that enables users to perform taxon-specific quantitation on large proteomic data sets without requiring high performance computing. This tool flexibly enables users to match the reference database settings to their experimental conditions, and can scale from two-organism studies to the entire UniProt repository. In addition, we provide a de novo analysis workflow that enables the identification of organisms in the sample without prior specification, analogous to 16S rRNA sequencing.

Enzymatic digestion for protein sequencing usually requires much time, and does not always result in high sequence coverage. Here we report the use of aqueous microdroplets to accelerate enzymatic reactions and, in particular, to improve protein sequencing. When a room temperature aqueous solution containing 10 µM myoglobin and 5 µg mL−1 trypsin is electrosonically sprayed (−3 kV) from a homemade setup to produce tiny (∼9 µm) microdroplets, we obtain 100% sequence coverage in less than 1 ms of digestion time, in sharp contrast to 60% coverage achieved by incubating the same solution at 37 °C for 14 h followed by analysis with a commercial electrospray ionization source that produces larger (∼60 µm) droplets. We also confirm the sequence of the therapeutic antibody trastuzumab (∼148 kDa), with a sequence coverage of 100% for light chains and 85% for heavy chains, demonstrating the practical utility of microdroplets in drug development., Mass spectrometry (MS)-based protein sequencing usually relies on in-solution proteolytic digestion, which is time-consuming and inefficient for certain proteins. Here, the authors achieve full protein sequence coverage in less than 1 ms by subjecting protein-protease mixtures to electrosonic spray ionization-MS.

Pioneering work by among others the groups of Hunt and Mann at the start of this century opened up the era of mass spectrometry based phosphoproteomics1-3. Through advances in sample preparation, phosphopeptide enrichment, mass spectrometric detection, peptide sequencing and dedicated database searches this field has matured and it is nowadays possible to monitor thousands of protein phosphorylation events qualitatively and quantitatively4-8. Such phosphoproteomics studies are important as protein phosphorylation is regulating nearly all biological processes and the deregulation of phosphorylation has been associated with the onset of various diseases9-15. Notwithstanding this maturation and broad acceptance of mass spectrometry based phosphoproteomics by the research community, the mass spectrometric (MS) analysis of phosphorylation remains demanding, mainly due to the often severe substoichiometric levels of phosphorylation and the intrinsic lability of the phosphate group, hampering both enrichment and unambiguous sequencing analysis of phosphopeptides. Here we review the increasing knowledge gathered about the mechanisms behind the fragmentation of phosphopeptide ions and describe additionally recent advances made to improve and facilitate the identification and site localization of phosphorylated peptides. Since our earlier related review in 200916 several fragmentation methods have been introduced and successfully applied to phosphopeptides, such as electron capture (ECD), electron transfer induced dissociation(ETD), and hybrid fragmentation techniques such as EThcD (a combination of ETD and HCD) and AI-ETD (a combination of ETD with infrared photo-activation). A plethora of data has been gathered on endogenous Ser, Thr, Tyr phosphorylation, but also on large synthetic phosphopeptide libraries, providing new insight into the mechanisms behind phosphopeptide fragmentation. Moreover, data on phosphopeptides harboring phosphorylated Histidine, Arginine or Lysine have become available, which exhibit some distinct fragmentation behaviors. Through all this acquired knowledge on phosphopeptide fragmentation and site localization mass spectrometry based phosphoproteomics has matured into a respectable tool for the broader life science research community.

RNA binding protein (RBP) plays an important role in cellular processes. Identifying RBPs by computation and experiment are both essential. Recently, an RBP predictor, RBPPred, is proposed in our group to predict RBPs. However, RBPPred is too slow for that it needs to generate PSSM matrix as its feature. Herein, based on the protein feature of RBPPred and Convolutional Neural Network (CNN), we develop a deep learning model called Deep-RBPPred. With the balance and imbalance training set, we obtain Deep-RBPPred-balance and Deep-RBPPred-imbalance models. Deep-RBPPred has three advantages comparing to previous methods. (1) Deep-RBPPred only needs few physicochemical properties based on protein sequences. (2) Deep-RBPPred runs much faster. (3) Deep-RBPPred has a good generalization ability. In the meantime, Deep-RBPPred is still as good as the state-of-the-art method. Testing in A. thaliana, S. cerevisiae and H. sapiens proteomes, MCC values are 0.82 (0.82), 0.65 (0.69) and 0.85 (0.80) for balance model (imbalance model) when the score cutoff is set to 0.5, respectively. In the same testing dataset, different machine learning algorithms (CNN and SVM) are also compared. The results show that CNN-based model can identify more RBPs than SVM-based. In comparing the balance and imbalance model, both CNN-base and SVM-based tend to favor the majority class in the imbalance set. Deep-RBPPred forecasts 280 (balance model) and 265 (imbalance model) of 299 new RBP. The sensitivity of balance model is about 7% higher than the state-of-the-art method. We also apply deep-RBPPred to 30 eukaryotes and 109 bacteria proteomes downloaded from Uniprot to estimate all possible RBPs. The estimating result shows that rates of RBPs in eukaryote proteomes are much higher than bacteria proteomes.

Protein-RNA complexes play important roles in various biological processes. The functions of protein-RNA complexes are dictated by their interactions, binding, stability, and affinity. In this work, we have identified the key residues (KRs), which are involved in both stability and binding. We found that 42 percent of considered proteins share common binding and stabilizing residues, whereas these residues are distinct in 58 percent of the proteins. Overall, 5 percent of stabilizing and 3 percent of binding residues serve as key residues. These residues are enriched with the combination of polar, charged, aliphatic, and aromatic residues. Analysis on subclasses of protein-RNA complexes based on protein structural class, function and RNA type showed that regulatory proteins, and complexes with single stranded RNA and rRNA have appreciable number of key residues. Specifically, Arg, Tyr, and Thr are preferred in most of the subclasses of protein-RNA complexes. In addition, residues with similar chemical behavior have different preferences to be KRs, such that Arg, Tyr, Val, and Thr are preferred over Lys, Trp, Ile, and Ser, respectively. Atomic level contacts revealed that charged and polar-nonpolar contacts are dominant in enzymes, polar in structural, and nonpolar in regulatory proteins. On the other hand, polar-nonpolar contacts are enriched in all these classes of protein-RNA complexes. Further, the influence of sequence and structural features such as conservation score, surrounding hydrophobicity, solvent accessibility, secondary structure, and long-range order in key residues are also discussed. We envisage that the present study provides insights to understand the structural and functional aspects of protein-RNA complexes.

Protein-Protein Interaction Sitesbase (PPInS), a high-performance database of protein-protein interacting interfaces, is presented. The atomic level information of the molecular interaction happening amongst various protein chains in protein-protein complexes (as reported in the Protein Data Bank [PDB]) together with their evolutionary information in Structural Classification of Proteins (SCOPe release 2.06), is made available in PPInS. Total 32468 PDB files representing X-ray crystallized multimeric protein-protein complexes with structural resolution better than 2.5 Å had been shortlisted to demarcate the protein-protein interaction interfaces (PPIIs). A total of 111857 PPIIs with ~32.24 million atomic contact pairs (ACPs) were generated and made available on a web server for on-site analysis and downloading purpose. All these PPIIs and protein-protein interacting patches (PPIPs) involved in them, were also analyzed in terms of a number of residues contributing in patch formation, their hydrophobic nature, amount of surface area they contributed in binding, and their homo and heterodimeric nature, to describe the diversity of information covered in PPInS. It was observed that 42.37% of total PPIPs were made up of 6–20 interacting residues, 53.08% PPIPs had interface area ≤1000 Å2 in PPII formation, 82.64% PPIPs were reported with hydrophobicity score of ≤10, and 73.26% PPIPs were homologous to each other with the sequence similarity score ranging from 75–100%. A subset “Non-Redundant Database (NRDB)” of the PPInS containing 2265 PPIIs, with over 1.8 million ACPs corresponding to the 1931 protein-protein complexes (PDBs), was also designed by removing structural redundancies at the level of SCOP superfamily (SCOP release 1.75). The web interface of the PPInS (http://www.cup.edu.in:99/ppins/home.php) offers an easy-to-navigate, intuitive and user-friendly environment, and can be accessed by providing PDB ID, SCOP superfamily ID, and protein sequence.

Motivation Template-based modeling, including homology modeling and protein threading, is a popular method for protein 3D structure prediction. However, alignment generation and template selection for protein sequences without close templates remain very challenging. Results We present a new method called DeepThreader to improve protein threading, including both alignment generation and template selection, by making use of deep learning (DL) and residue co-variation information. Our method first employs DL to predict inter-residue distance distribution from residue co-variation and sequential information (e.g. sequence profile and predicted secondary structure), and then builds sequence-template alignment by integrating predicted distance information and sequential features through an ADMM algorithm. Experimental results suggest that predicted inter-residue distance is helpful to both protein alignment and template selection especially for protein sequences without very close templates, and that our method outperforms currently popular homology modeling method HHpred and threading method CNFpred by a large margin and greatly outperforms the latest contact-assisted protein threading method EigenTHREADER. Availability and implementation http://raptorx.uchicago.edu/ Supplementary information Supplementary data are available at Bioinformatics online.

In this paper, we present a novel approach to solve the labeled point cloud superpositioning problem for performing structural comparisons of protein binding sites. The solution is based on a parallel evolution strategy that operates on large populations and runs on GPU hardware. The proposed evolution strategy reduces the likelihood of getting stuck in a local optimum of the multimodal real-valued optimization problem represented by labeled point cloud superpositioning. The performance of the GPU-based parallel evolution strategy is compared to a previously proposed CPU-based sequential approach for labeled point cloud superpositioning, indicating that the GPU-based parallel evolution strategy leads to qualitatively better results and significantly shorter runtimes, with speed improvements of up to a factor of 1,500 for large populations. Binary classification tests based on the ATP, NADH, and FAD protein subsets of CavBase, a database containing putative binding sites, show average classification rate improvements from about 92 percent (CPU) to 96 percent (GPU). Further experiments indicate that the proposed GPU-based labeled point cloud superpositioning approach can be superior to traditional protein comparison approaches based on sequence alignments.

Motivation ADP-ribosylation is a post-translational modification (PTM) implicated in several crucial cellular processes, ranging from regulation of DNA repair and chromatin structure to cell metabolism and stress responses. To date, a complete understanding of ADP-ribosylation targets and their modification sites in different tissues and disease states is still lacking. Identification of ADP-ribosylation sites is required to discern the molecular mechanisms regulated by this modification. This motivated us to develop a computational tool for the prediction of ADP-ribosylated sites. Results Here, we present ADPredict, the first dedicated computational tool for the prediction of ADP-ribosylated aspartic and glutamic acids. This predictive algorithm is based on (i) physicochemical properties, (ii) in-house designed secondary structure-related descriptors and (iii) three-dimensional features of a set of human ADP-ribosylated proteins that have been reported in the literature. ADPredict was developed using principal component analysis and machine learning techniques; its performance was evaluated both internally via intensive bootstrapping and in predicting two external experimental datasets. It outperformed the only other available ADP-ribosylation prediction tool, ModPred. Moreover, a novel secondary structure descriptor, HM-ratio, was introduced and successfully contributed to the model development, thus representing a promising tool for bioinformatics studies, such as PTM prediction. Availability and implementation ADPredict is freely available at www.ADPredict.net. Supplementary information Supplementary data are available at Bioinformatics online.

Aminoethylation of cysteines can provide enzymatically cleavable sites. The ability to obtain peptides containing antibody complementarity determining regions (CDRs) with aminoethylated cysteines was investigated. Because cysteines are often located N-terminal to CDRs, digestion with Lys-N enables acquisition of peptides with CDRs. Lys-N peptides containing an aminoethylated cysteine at the N-terminus were also amidinated. Subsequent collisional activation yields a unique loss of 118 Da that originates from this modified residue, providing a signature ion for cysteine-containing peptides. The relative cleavage efficiencies for Lys-N and trypsin are also compared.

Motivation The identification of signal peptides in protein sequences is an important step toward protein localization and function characterization. Results Here, we present DeepSig, an improved approach for signal peptide detection and cleavage-site prediction based on deep learning methods. Comparative benchmarks performed on an updated independent dataset of proteins show that DeepSig is the current best performing method, scoring better than other available state-of-the-art approaches on both signal peptide detection and precise cleavage-site identification. Availability and implementation DeepSig is available as both standalone program and web server at https://deepsig.biocomp.unibo.it. All datasets used in this study can be obtained from the same website. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Multiple sequence alignments (MSAs) can provide essential input to many bioinformatics applications, including protein structure prediction and functional annotation. However, the optimal selection of sequences to obtain biologically informative MSAs for such purposes is poorly explored, and has traditionally been performed manually. Results We present Selection of Alignment by Maximal Mutual Information (SAMMI), an automated, sequence-based approach to objectively select an optimal MSA from a large set of alternatives sampled from a general sequence database search. The hypothesis of this approach is that the mutual information among MSA columns will be maximal for those MSAs that contain the most diverse set possible of the most structurally and functionally homogeneous protein sequences. SAMMI was tested to select MSAs for functional site residue prediction by analysis of conservation patterns on a set of 435 proteins obtained from protein–ligand (peptides, nucleic acids and small substrates) and protein–protein interaction databases. Availability and implementation A freely accessible program, including source code, implementing SAMMI is available at https://github.com/nelsongil92/SAMMI.git. Supplementary information Supplementary data are available at Bioinformatics online.

Motivation A large number of protein sequences are becoming available through the application of novel high-throughput sequencing technologies. Experimental functional characterization of these proteins is time-consuming and expensive, and is often only done rigorously for few selected model organisms. Computational function prediction approaches have been suggested to fill this gap. The functions of proteins are classified using the Gene Ontology (GO), which contains over 40 000 classes. Additionally, proteins have multiple functions, making function prediction a large-scale, multi-class, multi-label problem. Results We have developed a novel method to predict protein function from sequence. We use deep learning to learn features from protein sequences as well as a cross-species protein–protein interaction network. Our approach specifically outputs information in the structure of the GO and utilizes the dependencies between GO classes as background information to construct a deep learning model. We evaluate our method using the standards established by the Computational Assessment of Function Annotation (CAFA) and demonstrate a significant improvement over baseline methods such as BLAST, in particular for predicting cellular locations. Availability and implementation Web server: http://deepgo.bio2vec.net, Source code: https://github.com/bio-ontology-research-group/deepgo Supplementary information Supplementary data are available at Bioinformatics online.

Motivation Protein–peptide interactions are one of the most important biological interactions and play crucial role in many diseases including cancer. Therefore, knowledge of these interactions provides invaluable insights into all cellular processes, functional mechanisms, and drug discovery. Protein–peptide interactions can be analyzed by studying the structures of protein–peptide complexes. However, only a small portion has known complex structures and experimental determination of protein–peptide interaction is costly and inefficient. Thus, predicting peptide-binding sites computationally will be useful to improve efficiency and cost effectiveness of experimental studies. Here, we established a machine learning method called SPRINT-Str (Structure-based prediction of protein–Peptide Residue-level Interaction) to use structural information for predicting protein–peptide binding residues. These predicted binding residues are then employed to infer the peptide-binding site by a clustering algorithm. Results SPRINT-Str achieves robust and consistent results for prediction of protein–peptide binding regions in terms of residues and sites. Matthews’ Correlation Coefficient (MCC) for 10-fold cross validation and independent test set are 0.27 and 0.293, respectively, as well as 0.775 and 0.782, respectively for area under the curve. The prediction outperforms other state-of-the-art methods, including our previously developed sequence-based method. A further spatial neighbor clustering of predicted binding residues leads to prediction of binding sites at 20–116% higher coverage than the next best method at all precision levels in the test set. The application of SPRINT-Str to protein binding with DNA, RNA and carbohydrate confirms the method‘s capability of separating peptide-binding sites from other functional sites. More importantly, similar performance in prediction of binding residues and sites is obtained when experimentally determined structures are replaced by unbound structures or quality model structures built from homologs, indicating its wide applicability. Availability and implementation http://sparks-lab.org/server/SPRINT-Str Supplementary information Supplementary data are available at Bioinformatics online.

Accurate classification on protein structural is playing an important role in Bioinformatics. An increase in evidence demonstrates that a variety of classification methods have been employed in such a field. In this research, the features of amino acids composition, secondary structure's feature, and correlation coefficient of amino acid dimers and amino acid triplets have been used. Flexible neutral tree (FNT), a particular tree structure neutral network, has been employed as the classification model in the protein structures' classification framework. Considering different feature groups owing diverse roles in the model, impact factors of different groups have been put forward in this research. In order to evaluate different impact factors, Impact Factors Scaling (IFS) algorithm, which aim at reducing redundant information of the selected features in some degree, have been put forward. To examine the performance of such framework, the 640, 1189, and ASTRAL datasets are employed as the low-homology protein structure benchmark datasets. Experimental results demonstrate that the performance of the proposed method is better than the other methods in the low-homology protein tertiary structures.

Motivation Computational methods for phosphorylation site prediction play important roles in protein function studies and experimental design. Most existing methods are based on feature extraction, which may result in incomplete or biased features. Deep learning as the cutting-edge machine learning method has the ability to automatically discover complex representations of phosphorylation patterns from the raw sequences, and hence it provides a powerful tool for improvement of phosphorylation site prediction. Results We present MusiteDeep, the first deep-learning framework for predicting general and kinase-specific phosphorylation sites. MusiteDeep takes raw sequence data as input and uses convolutional neural networks with a novel two-dimensional attention mechanism. It achieves over a 50% relative improvement in the area under the precision-recall curve in general phosphorylation site prediction and obtains competitive results in kinase-specific prediction compared to other well-known tools on the benchmark data. Availability and implementation MusiteDeep is provided as an open-source tool available at https://github.com/duolinwang/MusiteDeep. Supplementary information Supplementary data are available at Bioinformatics online.

The exponentially increasing number of protein and nucleic acid sequences provides opportunities to discover novel enzymes, metabolic pathways, and metabolites/natural products, thereby adding to our knowledge of biochemistry and biology. The challenge has evolved from generating sequence information to mining the databases to integrating and leveraging the available information, i.e., the availability of "genomic enzymology" web tools. Web tools that allow identification of biosynthetic gene clusters are widely used by the natural products/synthetic biology community, thereby facilitating the discovery of novel natural products and the enzymes responsible for their biosynthesis. However, many novel enzymes with interesting mechanisms participate in uncharacterized small-molecule metabolic pathways; their discovery and functional characterization also can be accomplished by leveraging information in protein and nucleic acid databases. This Perspective focuses on two genomic enzymology web tools that assist the discovery novel metabolic pathways: (1) Enzyme Function Initiative-Enzyme Similarity Tool (EFI-EST) for generating sequence similarity networks to visualize and analyze sequence-function space in protein families and (2) Enzyme Function Initiative-Genome Neighborhood Tool (EFI-GNT) for generating genome neighborhood networks to visualize and analyze the genome context in microbial and fungal genomes. Both tools have been adapted to other applications to facilitate target selection for enzyme discovery and functional characterization. As the natural products community has demonstrated, the enzymology community needs to embrace the essential role of web tools that allow the protein and genome sequence databases to be leveraged for novel insights into enzymological problems.

While peptide identifications in mass spectrometry (MS)-based shotgun proteomics are mostly obtained using database search methods, high-resolution spectrum data from modern MS instruments nowadays offer the prospect of improving the performance of computational de novo peptide sequencing. The major benefit of de novo sequencing is that it does not require a reference database to deduce full-length or partial tag-based peptide sequences directly from experimental tandem mass spectrometry spectra. Although various algorithms have been developed for automated de novo sequencing, the prediction accuracy of proposed solutions has been rarely evaluated in independent benchmarking studies. The main objective of this work is to provide a detailed evaluation on the performance of de novo sequencing algorithms on high-resolution data. For this purpose, we processed four experimental data sets acquired from different instrument types from collision-induced dissociation and higher energy collisional dissociation (HCD) fragmentation mode using the software packages Novor, PEAKS and PepNovo. Moreover, the accuracy of these algorithms is also tested on ground truth data based on simulated spectra generated from peak intensity prediction software. We found that Novor shows the overall best performance compared with PEAKS and PepNovo with respect to the accuracy of correct full peptide, tag-based and single-residue predictions. In addition, the same tool outpaced the commercial competitor PEAKS in terms of running time speedup by factors of around 12-17. Despite around 35% prediction accuracy for complete peptide sequences on HCD data sets, taken as a whole, the evaluated algorithms perform moderately on experimental data but show a significantly better performance on simulated data (up to 84% accuracy). Further, we describe the most frequently occurring de novo sequencing errors and evaluate the influence of missing fragment ion peaks and spectral noise on the accuracy. Finally, we discuss the potential of de novo sequencing for now becoming more widely used in the field.

De novo peptide sequencing using tandem mass spectrometry (MS/MS) data has become a major computational method for sequence identification in recent years. With the development of new instruments and technology, novel computational methods have emerged with enhanced performance. However, there are only a few methods focusing on ECD/ETD spectra, which mainly contain variants of $c$ -ions and $z$ -ions. Here, a de novo sequencing method for ECD/ETD spectra, NovoExD, is presented. NovoExD applies a new form of spectrum graph with multiple edge types (called a GMET), considers multiple peptide tags, and integrates amino acid combination (AAC) and fragment ion charge information. Its performance is compared with another successful de novo sequencing method, pNovo+, which has an option for ECD/ETD spectra. Experiments conducted on three different datasets show that the average full length peptide identification accuracy of NovoExD is as high as 88.70 percent, and that NovoExD's average accuracy is more than 20 percent greater on all datasets than that of pNovo+.

Motivation Antimicrobial peptides (AMPs) are innate immune molecules that exhibit activities against a range of microbes, including bacteria, fungi, viruses and protozoa. Recent increases in microbial resistance against current drugs has led to a concomitant increase in the need for novel antimicrobial agents. Over the last decade, a number of AMP prediction tools have been designed and made freely available online. These AMP prediction tools show potential to discriminate AMPs from non-AMPs, but the relative quality of the predictions produced by the various tools is difficult to quantify. Results We compiled two sets of AMP and non-AMP peptides, separated into three categories—antimicrobial, antibacterial and bacteriocins. Using these benchmark data sets, we carried out a systematic evaluation of ten publicly available AMP prediction methods. Among the six general AMP prediction tools—ADAM, CAMPR3(RF), CAMPR3(SVM), MLAMP, DBAASP and MLAMP—we find that CAMPR3(RF) provides a statistically significant improvement in performance, as measured by the area under the receiver operating characteristic (ROC) curve, relative to the other five methods. Surprisingly, for antibacterial prediction, the original AntiBP method significantly outperforms its successor, AntiBP2 based on one benchmark dataset. The two bacteriocin prediction tools, BAGEL3 and BACTIBASE, both provide very good performance and BAGEL3 outperforms its predecessor, BACTIBASE, on the larger of the two benchmarks. Supplementary information Supplementary data are available at Bioinformatics online.

BACKGROUND The immunogenicities of polypeptide blood group antigens vary widely. One possible determinant of immunogenicity is antigenic foreignness. The goal was to employ alternative ways of assessing foreignness and determine whether foreignness was related to immunogenicity. STUDY DESIGN AND METHODS Foreignness was assessed as the extent of protein functional disruption caused by the exofacial amino acid (AA) substitutions that create blood group antigens, using AA substitution prediction algorithms such as Meta-SNP and according to whether those substitutions were radical or conservative. RESULTS AA substitutions that create the most immunogenic antigens had the highest Meta-SNP scores, predictive of greater protein structure and function changes. Four of the 11 exofacial AAs that distinguish the most immunogenic antigen, RhD, from RhCE, and substitutions creating four of the five next most immunogenic antigens had the highest Meta-SNP scores (0.293-0.649). Excluding the outlier Jka, the mean Meta-SNP score of the four most immunogenic non-RhD antigens (K, Lua, E, c) was 3.7-fold higher than the mean of the four least immunogenic (M, Fya, C, S), 0.459 versus 0.123 (p = 0.0026). Regression analysis revealed a relationship between immunogenicity and Meta-SNP score (R2 = 0.953). Actual protein functional disruption was predicted for the AA substitution creating the E antigen. An AA cluster at Positions 350, 353, and 354 of RhD was unique, containing radical substitutions according to two classification schemes and relatively high Meta-SNP scores (0.351-0.432). CONCLUSION The immunogenicity of blood group antigens was related to the functional disruption caused by the AA substitutions that create the antigens, as measured by Meta-SNP score.

Background Proteins with low complexity regions (LCRs) have atypical sequence and structural features. Their amino acid composition varies from the expected, determined proteome-wise, and they do not follow the rules of structural folding that prevail in globular regions. One way to characterize these regions is by assessing the repeatability of a sequence, that is, calculating the local propensity of a region to be part of a repeat. Results We combine two local measures of low complexity, repeatability (using the RES algorithm) and fraction of the most frequent amino acid, to evaluate different proteomes, datasets of protein regions with specific features, and individual cases of proteins with extreme compositions. We apply a representation called ‘low complexity triangle’ as a proof-of-concept to represent the low complexity measured values. Results show that proteomes have distinct signatures in the low complexity triangle, and that these signatures are associated to complexity features of the sequences. We developed a web tool called LCT (http://cbdm-01.zdv.uni-mainz.de/~munoz/lct/) to allow users to calculate the low complexity triangle of a given protein or region of interest. Conclusions The low complexity triangle proves to be a suitable procedure to represent the general low complexity of a sequence or protein dataset. Homorepeats, direpeats, compositionally biased regions and globular regions occupy characteristic positions in the triangle. The described pipeline can be used to characterize LCRs and may help in quantifying the content of degenerated tandem repeats in proteins and proteomes.

Motivation Template-based modeling, the process of predicting the tertiary structure of a protein by using homologous protein structures, is useful if good templates can be found. Although modern homology detection methods can find remote homologs with high sensitivity, the accuracy of template-based models generated from homology-detection-based alignments is often lower than that from ideal alignments. Results In this study, we propose a new method that generates pairwise sequence alignments for more accurate template-based modeling. The proposed method trains a machine learning model using the structural alignment of known homologs. It is difficult to directly predict sequence alignments using machine learning. Thus, when calculating sequence alignments, instead of a fixed substitution matrix, this method dynamically predicts a substitution score from the trained model. We evaluate our method by carefully splitting the training and test datasets and comparing the predicted structure’s accuracy with that of state-of-the-art methods. Our method generates more accurate tertiary structure models than those produced from alignments obtained by other methods. Availability and implementation https://github.com/shuichiro-makigaki/exmachina. Supplementary information Supplementary data are available at Bioinformatics online.

Summary We report an update for the MAFFT multiple sequence alignment program to enable parallel calculation of large numbers of sequences. The G-INS-1 option of MAFFT was recently reported to have higher accuracy than other methods for large data, but this method has been impractical for most large-scale analyses, due to the requirement of large computational resources. We introduce a scalable variant, G-large-INS-1, which has equivalent accuracy to G-INS-1 and is applicable to 50 000 or more sequences. Availability and implementation This feature is available in MAFFT versions 7.355 or later at https://mafft.cbrc.jp/alignment/software/mpi.html. Supplementary information Supplementary data are available at Bioinformatics online.

Identification of functional elements for a protein of interest is important for achieving a mechanistic understanding. However, it remains cumbersome to assess each and every amino acid of a given protein in relevance to its functional significance. Here, we report a strategy, PArsing fragmented DNA Sequences from CRISPR Tiling MUtagenesis Screening (PASTMUS), which provides a streamlined workflow and a bioinformatics pipeline to identify critical amino acids of proteins in their native biological contexts. Using this approach, we map six proteins—three bacterial toxin receptors and three cancer drug targets, and acquire their corresponding functional maps at amino acid resolution.

With advancements in synthetic biology, the cost and the time needed for designing and synthesizing customized gene products have been steadily decreasing. Many research laboratories in academia as well as industry routinely create genetically engineered proteins as a part of their research activities. However, manipulation of protein sequences could result in unintentional production of toxic proteins. Therefore, being able to identify the toxicity of a protein before the synthesis would reduce the risk of potential hazards. Existing methods are too specific, which limits their application. Here, we extended general function prediction methods for predicting the toxicity of proteins. Protein function prediction methods have been actively studied in the bioinformatics community and have shown significant improvement over the last decade. We have previously developed successful function prediction methods, which were shown to be among top-performing methods in the community-wide functional annotation experiment, CAFA. Based on our function prediction method, we developed a neural network model, named NNTox, which uses predicted GO terms for a target protein to further predict the possibility of the protein being toxic. We have also developed a multi-label model, which can predict the specific toxicity type of the query sequence. Together, this work analyses the relationship between GO terms and protein toxicity and builds predictor models of protein toxicity.

Recent advances in DNA sequencing methods revolutionized biology by providing highly accurate reads, with high throughput or high read length. These read data are being used in many biological and medical applications. Modern DNA sequencing methods have no equivalent in protein sequencing, severely limiting the widespread application of protein data. Recently, several optical protein sequencing methods have been proposed that rely on the fluorescent labeling of amino acids. Here, we introduce the reprotonation-deprotonation protein sequencing method. Unlike other methods, this proposed technique relies on the measurement of an electrical signal and requires no fluorescent labeling. In reprotonation-deprotonation protein sequencing, the terminal amino acid is identified through its unique protonation signal, and by repeatedly cleaving the terminal amino acids one-by-one, each amino acid in the peptide is measured. By means of simulations, we show that, given a reference database of known proteins, reprotonation-deprotonation sequencing has the potential to correctly identify proteins in a sample. Our simulations provide target values for the signal-to-noise ratios that sensor devices need to attain in order to detect reprotonation-deprotonation events, as well as suitable pH values and required measurement times per amino acid. For instance, an SNR of 10 is required for a 61.71% proteome recovery rate with 100 ms measurement time per amino acid.

Machine learning (ML) is ubiquitous in bioinformatics, due to its versatility. One of the most crucial aspects to consider while training a ML model is to carefully select the optimal feature encoding for the problem at hand. Biophysical propensity scales are widely adopted in structural bioinformatics because they describe amino acids properties that are intuitively relevant for many structural and functional aspects of proteins, and are thus commonly used as input features for ML methods. In this paper we reproduce three classical structural bioinformatics prediction tasks to investigate the main assumptions about the use of propensity scales as input features for ML methods. We investigate their usefulness with different randomization experiments and we show that their effectiveness varies among the ML methods used and the tasks. We show that while linear methods are more dependent on the feature encoding, the specific biophysical meaning of the features is less relevant for non-linear methods. Moreover, we show that even among linear ML methods, the simpler one-hot encoding can surprisingly outperform the “biologically meaningful” scales. We also show that feature selection performed with non-linear ML methods may not be able to distinguish between randomized and “real” propensity scales by properly prioritizing to the latter. Finally, we show that learning problem-specific embeddings could be a simple, assumptions-free and optimal way to perform feature learning/engineering for structural bioinformatics tasks., SCOPUS: ar.j, info:eu-repo/semantics/published

BackgroundWhile DNA and RNA methods are routine to disrupt the expression of specific genes, complete understanding of developmental processes requires also protein methods, because: oocytes and early embryos accumulate proteins and these are not directly affected by DNA and RNA methods. When proteins in the oocyte encounter a specific antibody and theTRIpartiteMotiv-containing21(TRIM21) ubiquitin-protein ligase, they can be committed to degradation in the proteasome, producing a transient functional knock-out that reveals the role of the protein. However, there are doubts about whether this targeted proteolysis could be successfully used to study mammalian development, because duration of the transient effect is unknown, and also because amounts of reagents delivered must be adequate in relation to the amount of target protein, which is unknown, too.ResultsWe show that the mouse egg contains up to 1E-02 picomoles/protein, as estimated by mass spectrometry using the intensity-based absolute quantification (iBAQ) algorithm. However, the egg can only accommodate ≈1E-04 picomoles of antibody or TRIM21 without incurring toxic effects. Within this framework, we demonstrate that TRIM21-mediated protein depletion efficiently disrupts the embryonic process of trophectoderm formation, which critically depends on theTEA domain family member 4(Tead4) gene. TEAD4 depletion starting at the 1-cell stage lasts for 3 days prior to a return of gene and protein expression to baseline. This time period is long enough to result in a phenotype entirely consistent with that of the published null mutation and RNA interference studies: significant underexpression of trophectodermal genesCdx2andGata3and strongly impaired ability of embryos to cavitate and implant in the uterus. Omics data are available via ProteomeXchange (PXD012613) and GEO (GSE124844).ConclusionsTRIM21-mediated protein depletion can be an effective means to disrupt gene function in mouse development, provided the target gene is chosen carefully and the method is tuned accurately. The knowledge gathered in this study provides the basic know-how (prerequisites, requirements, limitations) to expedite the protein depletion of other genes besidesTead4.

Background HH-suite is a widely used open source software suite for sensitive sequence similarity searches and protein fold recognition. It is based on pairwise alignment of profile Hidden Markov models (HMMs), which represent multiple sequence alignments of homologous proteins. Results We developed a single-instruction multiple-data (SIMD) vectorized implementation of the Viterbi algorithm for profile HMM alignment and introduced various other speed-ups. These accelerated the search methods HHsearch by a factor 4 and HHblits by a factor 2 over the previous version 2.0.16. HHblits3 is ∼10× faster than PSI-BLAST and ∼20× faster than HMMER3. Jobs to perform HHsearch and HHblits searches with many query profile HMMs can be parallelized over cores and over cluster servers using OpenMP and message passing interface (MPI). The free, open-source, GPLv3-licensed software is available at https://github.com/soedinglab/hh-suite. Conclusion The added functionalities and increased speed of HHsearch and HHblits should facilitate their use in large-scale protein structure and function prediction, e.g. in metagenomics and genomics projects.

Proteases are multifunctional, promiscuous enzymes that degrade proteins as well as peptides and drive important processes in health and disease. Current technology has enabled the construction of libraries of peptide substrates that detect protease activity, which provides valuable biological information. An ideal library would be orthogonal, such that each protease only hydrolyzes one unique substrate, however this is impractical due to off-target promiscuity (i.e., one protease targets multiple different substrates). Therefore, when a library of probes is exposed to a cocktail of proteases, each protease activates multiple probes, producing a convoluted signature. Computational methods for parsing these signatures to estimate individual protease activities primarily use an extensive collection of all possible protease-substrate combinations, which require impractical amounts of training data when expanding to search for more candidate substrates. Here we provide a computational method for estimating protease activities efficiently by reducing the number of substrates and clustering proteases with similar cleavage activities into families. We envision that this method will be used to extract meaningful diagnostic information from biological samples., Author summary The activity of enzymatic proteins, which are called proteases, drives numerous important processes in health and disease: including cancer, immunity, and infectious disease. Many labs have developed useful diagnostics by designing sensors that measure the activity of these proteases. However, if we want to detect multiple proteases at the same time, it becomes impractical to design sensors that only detect one protease. This is due to a phenomenon called protease promiscuity, which means that proteases will activate multiple different sensors. Computational methods have been created to solve this problem, but the challenge is that these often require large amounts of training data. Further, completely different proteases may be detected by the same subset of sensors. In this work, we design a computational method to overcome this problem by clustering similar proteases into "subfamilies", which increases estimation accuracy. Further, our method tests multiple combinations of sensors to maintain accuracy while minimizing the number of sensors used. Together, we envision that this work will increase the amount of useful information we can extract from biological samples, which may lead to better clinical diagnostics.

Human whole-genome-sequencing reveals about 4 000 000 genomic variants per individual. These data are mostly stored as VCF-format files. Although many variant analysis methods accept VCF as input, many other tools require DNA or protein sequences, particularly for splicing prediction, sequence alignment, phylogenetic analysis, and structure prediction. However, there is no existing webserver capable of extracting DNA/protein sequences for genomic variants from VCF files in a user-friendly and efficient manner. We developed the SeqTailor webserver to bridge this gap, by enabling rapid extraction of (i) DNA sequences around genomic variants, with customizable window sizes and options to annotate the splice sites closest to the variants and to consider the neighboring variants within the window; and (ii) protein sequences encoded by the DNA sequences around genomic variants, with built-in SnpEff annotator and customizable window sizes. SeqTailor supports 11 species, including: human (GRCh37/GRCh38), chimpanzee, mouse, rat, cow, chicken, lizard, zebrafish, fruitfly, Arabidopsis and rice. Standalone programs are provided for command-line-based needs. SeqTailor streamlines the sequence extraction process, and accelerates the analysis of genomic variants with software requiring DNA/protein sequences. It will facilitate the study of genomic variation, by increasing the feasibility of sequence-based analysis and prediction. The SeqTailor webserver is freely available at http://shiva.rockefeller.edu/SeqTailor/.

Antibodies are essential to functional immunity, yet the epitopes targeted by antibody repertoires remain largely uncharacterized. To aid in characterization, we developed a generalizable strategy to predict antibody-binding epitopes within individual proteins and entire proteomes. Specifically, we selected antibody-binding peptides for 273 distinct sera out of a random library and identified the peptides using next-generation sequencing. To predict antibody-binding epitopes and the antigens from which these epitopes were derived, we tiled the sequences of candidate antigens into short overlapping subsequences of length k (k-mers). We used the enrichment over background of these k-mers in the antibody-binding peptide dataset to predict antibody-binding epitopes. As a positive control, we used this approach, termed K-mer Tiling of Protein Epitopes (K-TOPE), to predict epitopes targeted by monoclonal and polyclonal antibodies of well-characterized specificity, accurately recovering their known epitopes. K-TOPE characterized a commonly targeted antigen from Rhinovirus A, predicting four epitopes recognized by antibodies present in 87% of sera (n = 250). An analysis of 2,908 proteins from 400 viral taxa that infect humans predicted seven enterovirus epitopes and five Epstein-Barr virus epitopes recognized by >30% of specimens. Analysis of Staphylococcus and Streptococcus proteomes similarly predicted 22 epitopes recognized by >30% of specimens. Twelve of these common viral and bacterial epitopes agreed with previously mapped epitopes with p-values < 0.05. Additionally, we predicted 30 HSV2-specific epitopes that were 100% specific against HSV1 in novel and previously reported antigens. Experimentally validating these candidate epitopes could help identify diagnostic biomarkers, vaccine components, and therapeutic targets. The K-TOPE approach thus provides a powerful new tool to elucidate the organisms, antigens, and epitopes targeted by human antibody repertoires.

Transferring knowledge between species is key for many biological applications, but is complicated by divergent and convergent evolution. Many current approaches for this problem leverage sequence and interaction network data to transfer knowledge across species, exemplified by network alignment methods. While these techniques do well, they are limited in scope, creating metrics to address one specific problem or task. We take a different approach by creating an environment where multiple knowledge transfer tasks can be performed using the same protein representations. Specifically, our kernel-based method, MUNK, integrates sequence and network structure to create functional protein representations, embedding proteins from different species in the same vector space. First we show proteins in different species that are close in MUNK-space are functionally similar. Next, we use these representations to share knowledge of synthetic lethal interactions between species. Importantly, we find that the results using MUNK-representations are at least as accurate as existing algorithms for these tasks. Finally, we generalize the notion of a phenolog (‘orthologous phenotype’) to use functionally similar proteins (i.e. those with similar representations). We demonstrate the utility of this broadened notion by using it to identify known phenologs and novel non-obvious ones supported by current research.

Motivation Genetic variants in drug targets and metabolizing enzymes often have important functional implications, including altering the efficacy and toxicity of drugs. Identifying single nucleotide variants (SNVs) that contribute to differences in drug response and understanding their underlying mechanisms are fundamental to successful implementation of the precision medicine model. This work reports an effort to collect, classify and analyze SNVs that may affect the optimal response to currently approved drugs. Results An integrated approach was taken involving data mining across multiple information resources including databases containing drugs, drug targets, chemical structures, protein–ligand structure complexes, genetic and clinical variations as well as protein sequence alignment tools. We obtained 2640 SNVs of interest, most of which occur rarely in populations (minor allele frequency Availability and Implementation Data are available from Supplementary information file. Supplementary information Supplementary Tables S1–S5 are available at Bioinformatics online.