Superfamily Classification Research Articles

Testing the Capability of Embedding-Based Alignments on the GST Superfamily Classification: The Role of Protein Length.

In order to shed light on the usage of protein language model-based alignment procedures, we attempted the classification of Glutathione S-transferases (GST; EC 2.5.1.18) and compared our results with the ARBA/UNI rule-based annotation in UniProt. GST is a protein superfamily involved in cellular detoxification from harmful xenobiotics and endobiotics, widely distributed in prokaryotes and eukaryotes. What is particularly interesting is that the superfamily is characterized by different classes, comprising proteins from different taxa that can act in different cell locations (cytosolic, mitochondrial and microsomal compartments) with different folds and different levels of sequence identity with remote homologs. For this reason, GST functional annotation in a specific class is problematic: unless a structure is released, the protein can be classified only on the basis of sequence similarity, which excludes the annotation of remote homologs. Here, we adopt an embedding-based alignment to classify 15,061 GST proteins automatically annotated by the UniProt-ARBA/UNI rules. Embedding is based on the Meta ESM2-15b protein language. The embedding-based alignment reaches more than a 99% rate of perfect matching with the UniProt automatic procedure. Data analysis indicates that 46% of the UniProt automatically classified proteins do not conserve the typical length of canonical GSTs, whose structure is known. Therefore, 46% of the classified proteins do not conserve the template/s structure required for their family classification. Our approach finds that 41% of 64,207 GST UniProt proteins not yet assigned to any class can be classified consistently with the structural template length.

ContactLib-ATT: A Structure-Based Search Engine for Homologous Proteins.

General-purpose protein structure embedding can be used for many important protein biology tasks, such as protein design, drug design and binding affinity prediction. Recent researches have shown that attention-based encoder layers are more suitable to learn high-level features. Based on this key observation, we propose a two-level general-purpose protein structure embedding neural network, called ContactLib-ATT. On local embedding level, a biologically more meaningful contact context is introduced. On global embedding level, attention-based encoder layers are employed for better global representation learning. Our general-purpose protein structure embedding framework is trained and tested on the SCOP40 2.07 dataset. As a result, ContactLib-ATT achieves a SCOP superfamily classification accuracy of 82.4% (i.e., 6.7% higher than state-of-the-art method). On the same dataset, ContactLib-ATT is used to simulate a structure-based search engine for remote homologous proteins, and our top-10 candidate list contains at least one remote homolog with a probability of 91.9%.

AlphaFold2 reveals commonalities and novelties in protein structure space for 21 model organisms

Deep-learning (DL) methods like DeepMind’s AlphaFold2 (AF2) have led to substantial improvements in protein structure prediction. We analyse confident AF2 models from 21 model organisms using a new classification protocol (CATH-Assign) which exploits novel DL methods for structural comparison and classification. Of ~370,000 confident models, 92% can be assigned to 3253 superfamilies in our CATH domain superfamily classification. The remaining cluster into 2367 putative novel superfamilies. Detailed manual analysis on 618 of these, having at least one human relative, reveal extremely remote homologies and further unusual features. Only 25 novel superfamilies could be confirmed. Although most models map to existing superfamilies, AF2 domains expand CATH by 67% and increases the number of unique ‘global’ folds by 36% and will provide valuable insights on structure function relationships. CATH-Assign will harness the huge expansion in structural data provided by DeepMind to rationalise evolutionary changes driving functional divergence.

Isofunctional Clustering and Conformational Analysis of the Arsenate Reductase Superfamily Reveals Nine Distinct Clusters.

Arsenate reductase (ArsC) is a superfamily of enzymes that reduce arsenate. Due to active site similarities, some ArsC can function as low-molecular weight protein tyrosine phosphatases (LMW-PTPs). Broad superfamily classifications align with redox partners (Trx- or Grx-linked). To understand this superfamily's mechanistic diversity, the ArsC superfamily is classified on the basis of active site features utilizing the tools TuLIP (two-level iterative clustering process) and autoMISST (automated multilevel iterative sequence searching technique). This approach identified nine functionally relevant (perhaps isofunctional) protein groups. Five groups exhibit distinct ArsC mechanisms. Three are Grx-linked: group 4AA (classical ArsC), group 3AAA (YffB-like), and group 5BAA. Two are Trx-linked: groups 6AAAAA and 7AAAAAAAA. One is an Spx-like transcriptional regulatory group, group 5AAA. Three are potential LMW-PTP groups: groups 7BAAAA, and 7AAAABAA, which have not been previously identified, and the well-studied LMW-PTP family group 8AAA. Molecular dynamics simulations were utilized to explore functional site details. In several families, we confirm and add detail to literature-based mechanistic information. Mechanistic roles are hypothesized for conserved active site residues in several families. In three families, simulations of the unliganded structure sample specific conformational ensembles, which are proposed to represent either a more ligand-binding-competent conformation or a pathway toward a more binding-competent state; these active sites may be designed to traverse high-energy barriers to the lower-energy conformations necessary to more readily bind ligands. This more detailed biochemical understanding of ArsC and ArsC-like PTP mechanisms opens possibilities for further understanding of arsenate bioremediation and the LMW-PTP mechanism.

Transcriptional characterisation of the Exaiptasia pallida pedal disc

BackgroundBiological adhesion (bioadhesion), enables organisms to attach to surfaces as well as to a range of other targets. Bioadhesion evolved numerous times independently and is ubiquitous throughout the kingdoms of life. To date, investigations have focussed on various taxa of animals, plants and bacteria, but the fundamental processes underlying bioadhesion and the degree of conservation in different biological systems remain poorly understood. This study had two aims: 1) To characterise tissue-specific gene regulation in the pedal disc of the model cnidarian Exaiptasia pallida, and 2) to elucidate putative genes involved in pedal disc adhesion.ResultsFive hundred and forty-seven genes were differentially expressed in the pedal disc compared to the rest of the animal. Four hundred and twenty-seven genes were significantly upregulated and 120 genes were significantly downregulated. Forty-one condensed gene ontology terms and 19 protein superfamily classifications were enriched in the pedal disc. Eight condensed gene ontology terms and 11 protein superfamily classifications were depleted. Enriched superfamilies were consistent with classifications identified previously as important for the bioadhesion of unrelated marine invertebrates. A host of genes involved in regulation of extracellular matrix generation and degradation were identified, as well as others related to development and immunity. Ab initio prediction identified 173 upregulated genes that putatively code for extracellularly secreted proteins.ConclusionThe analytical workflow facilitated identification of genes putatively involved in adhesion, immunity, defence and development of the E. pallida pedal disc. When defence, immunity and development-related genes were identified, those remaining corresponded most closely to formation of the extracellular matrix (ECM), implicating ECM in the adhesion of anemones to surfaces. This study therefore provides a valuable high-throughput resource for the bioadhesion community and lays a foundation for further targeted research to elucidate bioadhesion in the Cnidaria.

Disjoint motif discovery in biological network using pattern join method.

The biological network plays a key role in protein function annotation, protein superfamily classification, disease diagnosis, etc. These networks exhibit global properties like small‐world property, power‐law degree distribution, hierarchical modularity, robustness, etc. Along with these, the biological network also possesses some local properties like clustering and network motif. Network motifs are recurrent and statistically over‐represented subgraphs in a target network. Operation of a biological network is controlled by these motifs, and they are responsible for many biological applications. Discovery of network motifs is a computationally hard problem and involves a subgraph isomorphism check which is NP‐complete. In recent years, researchers have developed various tools and algorithms to detect network motifs efficiently. However, it is still a challenging task to discover the network motif within a practical time bound for the large motif. In this study, an efficient pattern‐join based algorithm is proposed to discover network motif in biological networks. The performance of the proposed algorithm is evaluated on the transcription regulatory network of Escherichia coli and the protein interaction network of Saccharomyces cerevisiae. The running time of the proposed algorithm outperforms most of the existing algorithms to discover large motifs.

Complete Genome Sequence of Serratia marcescens Phage MTx.

Serratia marcescens is a nosocomial pathogen that has evolved resistance to multiple antibiotics. Here, we present the genome sequence of myophage MTx that infects S. marcescens MTx encodes 103 proteins, with 26 being assigned a predicted function or superfamily classification, and it has little similarity with other phages at the nucleotide level.

The Short-chain Dehydrogenase/Reductase Engineering Database (SDRED): A classification and analysis system for a highly diverse enzyme family.

The Short-chain Dehydrogenases/Reductases Engineering Database (SDRED) covers one of the largest known protein families (168 150 proteins). Assignment to the superfamilies of Classical and Extended SDRs was achieved by global sequence similarity and by identification of family-specific sequence motifs. Two standard numbering schemes were established for Classical and Extended SDRs that allow for the determination of conserved amino acid residues, such as cofactor specificity determining positions or superfamily specific sequence motifs. The comprehensive sequence dataset of the SDRED facilitates the refinement of family-specific sequence motifs. The glycine-rich motifs for Classical and Extended SDRs were refined to improve the precision of superfamily classification. In each superfamily, the majority of sequences formed a tightly connected sequence network and belonged to a large homologous family. Despite their different sequence motifs and their different sequence length, the two sequence networks of Classical and Extended SDRs are not separate, but connected by edges at a threshold of 40% sequence similarity, indicating that all SDRs belong to a large, connected network. The SDRED is accessible at https://sdred.biocatnet.de/.

A secondary cut­off threshold for improved HMMERCTTER protein superfamily classification

A secondary cut­off threshold for improved HMMERCTTER protein superfamily classification

ProtozoaDB 2.0: A Trypanosoma Brucei Case Study.

Over the last decade new species of Protozoa have been sequenced and deposited in GenBank. Analyzing large amounts of genomic data, especially using Next Generation Sequencing (NGS), is not a trivial task, considering that researchers used to deal or focus their studies on few genes or gene families or even small genomes. To facilitate the information extraction process from genomic data, we developed a database system called ProtozoaDB that included five genomes of Protozoa in its first version. In the present study, we present a new version of ProtozoaDB called ProtozoaDB 2.0, now with the genomes of 22 pathogenic Protozoa. The system has been fully remodeled to allow for new tools and a more expanded view of data, and now includes a number of analyses such as: (i) similarities with other databases (model organisms, the Conserved Domains Database, and the Protein Data Bank); (ii) visualization of KEGG metabolic pathways; (iii) the protein structure from PDB; (iv) homology inferences; (v) the search for related publications in PubMed; (vi) superfamily classification; and (vii) phenotype inferences based on comparisons with model organisms. ProtozoaDB 2.0 supports RESTful Web Services to make data access easier. Those services were written in Ruby language using Ruby on Rails (RoR). This new version also allows a more detailed analysis of the object of study, as well as expanding the number of genomes and proteomes available to the scientific community. In our case study, a group of prenyltransferase proteinsalready described in the literature was found to be a good drug target for Trypanosomatids.

SCOPe: Manual Curation and Artifact Removal in the Structural Classification of Proteins – extended Database

SCOPe (Structural Classification of Proteins—extended, http://scop.berkeley.edu) is a database of relationships between protein structures that extends the Structural Classification of Proteins (SCOP) database. SCOP is an expert-curated ordering of domains from the majority of proteins of known structure in a hierarchy according to structural and evolutionary relationships. SCOPe classifies the majority of protein structures released since SCOP development concluded in 2009, using a combination of manual curation and highly precise automated tools, aiming to have the same accuracy as fully hand-curated SCOP releases. SCOPe also incorporates and updates the ASTRAL compendium, which provides several databases and tools to aid in the analysis of the sequences and structures of proteins classified in SCOPe. SCOPe continues high-quality manual classification of new superfamilies, a key feature of SCOP. Artifacts such as expression tags are now separated into their own class, in order to distinguish them from the homology-based annotations in the remainder of the SCOPe hierarchy. SCOPe 2.06 contains 77,439 Protein Data Bank entries, double the 38,221 structures classified in SCOP.

Functional classification of CATH superfamilies: a domain-based approach for protein function annotation.

Functional classification of CATH superfamilies: a domain-based approach for protein function annotation.

Conotoxins: Structure, Therapeutic Potential and Pharmacological Applications.

Cone snails, also known as marine gastropods, from Conus genus produce in their venom a diverse range of small pharmacologically active structured peptides called conotoxins. The cone snail venoms are widely unexplored arsenal of toxins with therapeutic and pharmacological potential, making them a treasure trove of ligands and peptidic drug leads. Conotoxins are small disulfide bonded peptides, which act as remarkable selective inhibitors and modulators of ion channels (calcium, sodium, potassium), nicotinic acetylcholine receptors, noradrenaline transporters, N-methyl-D-aspartate receptors, and neurotensin receptors. They are highly potent and specific against several neuronal targets making them valuable as research tools, drug leads and even therapeutics. In this review, we discuss their gene superfamily classification, nomenclature, post-translational modification, structural framework, pharmacology and medical applications of the active conopeptides. We aim to give an overview of their structure and therapeutic potential. Understanding these aspects of conopeptides will help in designing more specific peptidic analogues.

Automatic Classification of Protein Structure Using the Maximum Contact Map Overlap Metric

In this work, we propose a new distance measure for comparing two protein structures based on their contact map representations. We show that our novel measure, which we refer to as the maximum contact map overlap (max-CMO) metric, satisfies all properties of a metric on the space of protein representations. Having a metric in that space allows one to avoid pairwise comparisons on the entire database and, thus, to significantly accelerate exploring the protein space compared to no-metric spaces. We show on a gold standard superfamily classification benchmark set of 6759 proteins that our exact k-nearest neighbor (k-NN) scheme classifies up to 224 out of 236 queries correctly and on a larger, extended version of the benchmark with 60; 850 additional structures, up to 1361 out of 1369 queries. Our k-NN classification thus provides a promising approach for the automatic classification of protein structures based on flexible contact map overlap alignments.

Functional classification of CATH superfamilies: a domain-based approach for protein function annotation

Computational approaches that can predict protein functions are essential to bridge the widening function annotation gap especially since <1.0% of all proteins in UniProtKB have been experimentally characterised. We present a domain-based method for protein function classification and prediction of functional sites that exploits functional subclassification of CATH superfamilies. The superfamilies are subclassified into functional families (FunFams) using a hierarchical clustering algorithm supervised by a new classification method, FunFHMMer.

Functional classification of CATH superfamilies: a domain-based approach for protein function annotation.

Motivation: Computational approaches that can predict protein functions are essential to bridge the widening function annotation gap especially since <1.0% of all proteins in UniProtKB have been experimentally characterized. We present a domain-based method for protein function classification and prediction of functional sites that exploits functional sub-classification of CATH superfamilies. The superfamilies are sub-classified into functional families (FunFams) using a hierarchical clustering algorithm supervised by a new classification method, FunFHMMer.Results: FunFHMMer generates more functionally coherent groupings of protein sequences than other domain-based protein classifications. This has been validated using known functional information. The conserved positions predicted by the FunFams are also found to be enriched in known functional residues. Moreover, the functional annotations provided by the FunFams are found to be more precise than other domain-based resources. FunFHMMer currently identifies 110 439 FunFams in 2735 superfamilies which can be used to functionally annotate > 16 million domain sequences.Availability and implementation: All FunFam annotation data are made available through the CATH webpages (http://www.cathdb.info). The FunFHMMer webserver (http://www.cathdb.info/search/by_funfhmmer) allows users to submit query sequences for assignment to a CATH FunFam.Contact: sayoni.das.12@ucl.ac.ukSupplementary information: Supplementary data are available at Bioinformatics online.

Sequence-based protein superfamily classification using computational intelligence techniques: a review.

Protein superfamily classification deals with the problem of predicting the family membership of newly discovered amino acid sequence. Although many trivial alignment methods are already developed by previous researchers, but the present trend demands the application of computational intelligent techniques. As there is an exponential growth in size of biological database, retrieval and inference of essential knowledge in the biological domain become a very cumbersome task. This problem can be easily handled using intelligent techniques due to their ability of tolerance for imprecision, uncertainty, approximate reasoning, and partial truth. This paper discusses the various global and local features extracted from full length protein sequence which are used for the approximation and generalisation of the classifier. The various parameters used for evaluating the performance of the classifiers are also discussed. Therefore, this review article can show right directions to the present researchers to make an improvement over the existing methods.

Structural templates of disordered granular media

Granular materials, in common with many complex systems, exhibit a range of self-organization processes that control their mechanical performance. Many of these processes directly manifest in the evolution of the contact network as the material responds to applied stresses and strains. Yet the connections between the topology, structure and dynamics of this evolving contact network remain poorly understood. Here we demonstrate that dense granular systems under a variety of loading conditions exhibit preferred structural ordering reminiscent of a superfamily classification. In particular, two distinct superfamilies are discovered: the first is typically exhibited by materials in the pre-failure regime, while the second manifests in the unstable or failure regime. We demonstrate the robustness of these findings with respect to a range of packing fractions in experimental sand and photoelastic disk assemblies subject to compression and shear, as well as in a series of discrete element simulations of compression tests. We show that the superfamily classification of small connected subgraphs in a granular material can be used to map boundaries in a so-called jamming phase diagram and, consequently, offers a key opportunity to bridge the mechanics and physics perspectives on the constitutive behavior of granular systems.

Clustering HLA class I superfamilies using structural interaction patterns.

Human leukocyte antigen (HLA) class I molecules are critical components of the cell-mediated immune system that bind and present intracellular antigenic peptides to CD8+ T cell receptors. To understand the interaction mechanism underlying human leukocyte antigen (HLA) class I specificity in detail, we studied the structural interaction characteristics of 16,393 nonameric peptides binding to 58 HLA-A and -B molecules. Our analysis showed for the first time that HLA-peptide intermolecular bonding patterns vary among different alleles and may be grouped in a superfamily dependent manner. Through the use of these HLA class I ‘fingerprints’, a high resolution HLA class I superfamily classification schema was developed. This classification is capable of separating HLA alleles into well resolved, non-overlapping clusters, which is consistent with known HLA superfamily definitions. Such structural interaction approach serves as an excellent alternative to the traditional methods of HLA superfamily definitions that use peptide binding motifs or receptor information, and will help identify appropriate antigens suitable for broad-based subunit vaccine design.

Two-Stage Approach for Protein Superfamily Classification

We deal with the problem of protein superfamily classification in which the family membership of newly discovered amino acid sequence is predicted. Correct prediction is a matter of great concern for the researchers and drug analyst which helps them in discovery of new drugs. As this problem falls broadly under the category of pattern classification problem, we have made all efforts to optimize feature extraction in the first stage and classifier design in the second stage with an overall objective to maximize the performance accuracy of the classifier. In the feature extraction phase, Genetic Algorithm- (GA-) based wrapper approach is used to select few eigenvectors from the principal component analysis (PCA) space which are encoded as binary strings in the chromosome. On the basis of position of 1’s in the chromosome, the eigenvectors are selected to build the transformation matrix which then maps the original high-dimension feature space to lower dimension feature space. Using PCA-NSGA-II (non-dominated sorting GA), the nondominated solutions obtained from the Pareto front solve the trade-off problem by compromising between the number of eigenvectors selected and the accuracy obtained by the classifier. In the second stage, recursive orthogonal least square algorithm (ROLSA) is used for training radial basis function network (RBFN) to select optimal number of hidden centres as well as update the output layer weighting matrix. This approach can be applied to large data set with much lower requirements of computer memory. Thus, very small architectures having few number of hidden centres are obtained showing higher level of performance accuracy.