TIMELESS protein (TIM) protects replication forks from stalling at difficult-to-replicate regions and plays an important role in DNA damage response, including checkpoint signaling, protection of stalled replication forks and DNA repair. Loss of TIM causes severe replication stress, while its overexpression is common in various types of cancer, providing protection from DNA damage and resistance to chemotherapy. Although TIM has mostly been studied for its part in replication stress response, its additional roles in supporting genome stability and a wide variety of other cellular pathways are gradually coming to light. This review discusses the diverse functions of TIM and its orthologs in healthy and cancer cells, open questions, and potential future directions.

Telomeres and their single stranded overhangs gradually shorten with successive cell divisions, as part of the natural aging process, but can be elongated by telomerase, a nucleoprotein complex which is activated in the majority of cancers. This prominent implication in cancer and aging has made the repetitive telomeric sequences (TTAGGG repeats) and the G-quadruplex structures that form in their overhangs the focus of intense research in the past several decades. However, until recently most in vitro efforts to understand the structure, stability, dynamics, and interactions of telomeric overhangs had been focused on short sequences that are not representative of longer sequences encountered in a physiological setting. In this review, we will provide a broad perspective about telomeres and associated factors, and introduce the agents and structural characteristics involved in organizing, maintaining, and protecting telomeric DNA. We will also present a summary of recent research performed on long telomeric sequences, nominally defined as those that can form two or more tandem G-quadruplexes, i.e., which contain eight or more TTAGGG repeats. Results of experimental studies using a broad array of experimental tools, in addition to recent computational efforts will be discussed, particularly in terms of their implications for the stability, folding topology, and compactness of the tandem G-quadruplexes that form in long telomeric overhangs.

Alternative DNA structures that differ from the canonical B-form of DNA can arise from repetitive sequences and play beneficial roles in many cellular processes such as gene regulation and chromatin organization. However, they also threaten genomic stability in several ways including mutagenesis and collisions with replication and/or transcription machinery, which lead to genomic instability that is associated with human disease. Thus, the careful regulation of non-B-DNA structure formation and resolution is crucial for the maintenance of genome integrity. Several protein factors have been demonstrated to associate with alternative DNA structures to facilitate their removal, one of which is the ADP-ribose transferase (ART) PARP1 (also called ADP-ribosyltransferase diphtheria toxin-like 1 or ARTD1), a multifaceted DNA repair enzyme that recognizes single- and double-stranded DNA breaks and synthesizes chains of poly (ADP-ribose) (PAR) to recruit DNA repair proteins. It is now well appreciated that PARP1 recognizes several nucleic acid structures beyond DNA lesions, including stalled replication forks, DNA hairpins and cruciforms, R-loops, and DNA G-quadruplexes (G4 DNA). In this review, we summarize the current evidence of a direct association of PARP1 with each of these aforementioned alternative DNA structures, as well as discuss the role of PARP1 in the prevention of non-B-DNA structure-induced genetic instability. We will focus on the mechanisms of the recognition and binding by PARP1 to each alternative structure and the structure-based stimulation of PARP1 catalytic activity upon binding. Finally, we will discuss some of the outstanding gaps in the literature and offer speculative insight for questions that remain to be experimentally addressed.

The human methyltransferase MLL4 plays a critical role in embryogenesis and development, and aberrant activity of MLL4 is linked to neurodegenerative and developmental disorders and cancer. MLL4 contains the catalytic SET domain that catalyzes mono methylation of lysine 4 of histone H3 (H3K4me1) and seven plant homeodomain (PHD) fingers, six of which have not been structurally and functionally characterized. Here, we demonstrate that the triple PHD finger cassette of MLL4, harboring its fourth, fifth and sixth PHD fingers (MLL4PHD456) forms an integrated module, maintains the binding selectivity of the PHD6 finger toward acetylated lysine 16 of histone H4 (H4K16ac), and is capable of binding to DNA. Our findings highlight functional correlation between H4K16ac and H3K4me1, two major histone modifications that are recognized and written, respectively, by MLL4.

In early 1990s, several proteins were shown to depend on additional stretches of polypeptide (termed as prosequence/prodomain) for their folding. These regions of the protein were often termed as IMCs (Intra Molecular Chaperones), since they would be cleaved from the mature folded protein eventually. Such proteins were hypothesized to face a kinetic barrier to their folding, which was probably lowered by the prosequences. In last three decades, numerous examples of such proteins have accumulated in literature. Yet, no study has been reported so far attempting to understand the evolutionary differences and similaritess of such proteins. Till date such proteins are continued to be treated as anomalous variants, rather than as representatives of any alternate protein folding strategy. Do such proteins have any distinctive structural facets OR typical biological roles, necessitating an unconventional strategy of protein folding? Do prosequences carry any unique or conserved features that are essential to their function? ProSeqAProDb: ProSequence Assisted Protein Database, (which can be accessed at https://proseqaprodb.mkulab.in) was built as a comprehensive database, to systematically study such proteins along with their pro-sequences. The database currently contains 2140 prosequence assisted proteins (1848 eukaryotic, 255 bacterial, 24 viral and 13 archaeal proteins), from 690 organisms later categorised into 960 families. We envisage that the availability of this curated dataset will enable the researchers worldwide to further their investigation in the origin, importance and evolution of such proteins, leading to better understanding of the protein folding process as a whole.

Spatial proteomics aims for a global description of organelle-specific protein distribution and dynamics, which is essential for understanding the molecular functions and cellular processes in health and disease. However, the application of this technique is seriously restricted by the neglect of robustness among proteomic signatures identified using standard statistical frameworks. Moreover, it is still a major bottleneck to automatically interpretate the identified proteomic signatures due to lack of integration of subcellular information. Herein, an online-tool SISPRO was constructed to (a) identify proteomic signatures with good robustness and accuracy via collectively evaluating relative weighted consistency (CWrel) & area under the curve (AUC) and (b) interpretate the identified signature based on comprehensive subcellular information from 9 organelles and 22 subcellular structures. All in all, SISPRO provides the endeavor to realize the simultaneous improvement of robustness and accuracy in signature identification and the unique capacity in biological annotation lies in its wide coverage of proteins and comprehensive spatial information. SISPRO is expected to be critical in spatial proteomic studies, which can be freely accessed without any login requirement at https://idrblab.org/sispro/.

Although nascent RNA profiling data are widely used in transcriptional regulation studies, the development and standardization of data processing pipeline lags far behind RNA-seq. We are filling this gap by establishing the nASAP web server (https://grobase.top/nasap/) to provide practical quality evaluation and comprehensive analysis of nascent RNA datasets. In nASAP, four customized analysis modules are provided, including i) quality assessment, which summarizes the sequencing statistics, mapping ratio, and evaluates RNA integrity and mRNA contamination; ii) quantification analysis for mRNAs, lncRNAs and eRNAs; iii) pausing analysis across the whole genome based on sequencing reads distribution; and iv) network analysis to better understand the gene regulatory mechanism by obtaining annotated enhancer-promoter interactomes. The nASAP is user-friendly and outperforms the existing pipeline for quality control of nascent RNA profiling data. We anticipate that nASAP, which eases both basic and advanced analysis of nascent RNA data, will be extremely useful in various fields.

Conservation of gene neighbourhood over evolutionary distances is generally indicative of shared regulation or functional association among genes. This concept has been broadly exploited in prokaryotes but its use on eukaryotic genomes has been limited to specific functional classes, such as biosynthetic gene clusters. We here used an evolutionary-based gene cluster discovery algorithm (EvolClust) to pre-compute evolutionarily conserved gene neighbourhoods, which can be searched, browsed and downloaded in EvolClustDB. We inferred ∼35,000 cluster families in 882 different species in genome comparisons of five taxonomically broad clades: Fungi, Plants, Metazoans, Insects and Protists. EvolClustDB allows browsing through the cluster families, as well as searching by protein, species, identifier or sequence. Visualization allows inspecting gene order per species in a phylogenetic context, so that relevant evolutionary events such as gain, loss or transfer, can be inferred. EvolClustDB is freely available, without registration, at http://evolclustdb.org/.

An increasingly common output arising from the analysis of shotgun metagenomic datasets is the generation of metagenome-assembled genomes (MAGs), with tens of thousands of MAGs now described in the literature. However, the discovery and comparison of these MAG collections is hampered by the lack of uniformity in their generation, annotation and storage. To address this, we have developed MGnify Genomes, a growing collection of biome-specific non-redundant microbial genome catalogues generated using MAGs and publicly available isolate genomes. Genomes within a biome-specific catalogue are organised into species clusters. For species that contain multiple conspecific genomes, the highest quality genome is selected as the representative, always prioritising an isolate genome over a MAG. The species representative sequences and annotations can be visualised on the MGnify website and the full catalogue and associated analysis outputs can be downloaded from MGnify servers. A suite of online search tools is provided allowing users to compare their own sequences, ranging from a gene to sets of genomes, against the catalogues. Seven biomes are available currently, comprising over 300,000 genomes that represent 11,048 non-redundant species, and include 36 taxonomic classes not currently represented by cultured genomes. MGnify Genomes is available at https://www.ebi.ac.uk/metagenomics/browse/genomes/.

Combining multiple binding profiles, such as transcription factors and histone modifications, is a crucial step in revealing the functions of complex biological systems. Although a massive amount of chromatin immunoprecipitation followed by sequencing (ChIP-seq) data is available, existing ChIP-seq databases or repositories focus on individual experiments, and it is difficult to elucidate orchestrated regulation by DNA-binding elements. We developed the Comprehensive Collection and Comparison for ChIP-Seq Database (C4S DB) to provide researchers with insights into the combination of DNA binding elements based on quality-assessed public ChIP-seq data. The C4S DB is based on >16,000 human ChIP-seq experiments and provides two main web interfaces to discover the relationships between ChIP-seq data. "Gene browser" illustrates the landscape of distributions of binding elements around a specified gene, and "global similarity," a hierarchical clustering heatmap based on a similarity between two ChIP-seq experiments, gives an overview of genome-wide relations of regulatory elements. These functions promote the identification or evaluation of both gene-specific and genome-wide colocalization or mutually exclusive localization. Modern web technologies allow users to search for and aggregate large-scale experimental data through interactive web interfaces with quick responses. The C4S DB is available at https://c4s.site.

We present RNASequest, a customizable RNA sequencing (RNAseq) analysis, app management, and result publishing framework. Its three-in-one RNAseq data analysis ecosystem consists of (1) a reproducible, configurable expression analysis (EA) module, (2) multi-faceted result presentation in R Shiny, a Bookdown document and an online slide deck, and (3) a centralized data management system. In principle, following up our well-received omics data visualization tool Quickomics, RNASequest automates the differential gene expression analysis step, eases statistical model design by built-in covariates testing module, and further provides a web-based tool, ShinyOne, to manage apps powered by Quickomics and reports generated by running the pipeline on multiple projects in one place. Researchers can experience the functionalities by exploring demo data sets hosted at http://shinyone.bxgenomics.com or following the tutorial, https://interactivereport.github.io/RNASequest/tutorial/docs/introduction.html to set up the framework locally to process private RNAseq datasets. The source code released under MIT open-source license is provided at https://github.com/interactivereport/RNASequest.

Massive sequencing of microbiomes has led to the discovery of a large number of phage genomes with intermittent stop codon recoding. We have developed a computational tool, MgCod, that identifies genomic regions (blocks) with distinct stop codon recoding simultaneously with the prediction of protein-coding regions. When MgCod was used to scan a large volume of human metagenomic contigs hundreds of viral contigs with intermittent stop codon recoding were revealed. Many of these contigs originated from genomes of known crAssphages. Further analyses had shown that intermittent recoding was associated with subtle patterns in the organization of protein-coding genes, such as 'single-coding' and 'dual-coding'. The dual-coding genes, clustered into blocks, could be translated by two alternative codes producing nearly identical proteins. It was observed that the dual-coded blocks were enriched with the early-stage phage genes, while the late-stage genes were residing in the single-coded blocks. MgCod can identify types of stop codon recoding in novel genomic sequences in parallel with gene prediction. It is available for download from https://github.com/gatech-genemark/MgCod.

The multiple sequence alignment (MSA) is the entry point of many RNA structure modeling tasks, such as prediction of RNA secondary structure (rSS) and contacts. However, there are few automated programs for generating high quality MSAs of target RNA molecules. We have developed rMSA, a hierarchical pipeline for sensitive search and accurate alignment of RNA homologs for a target RNA. On a diverse set of 365 non-redundant RNA structures, rMSA significantly outperforms an existing MSA generation method (RNAcmap) by approximately 20% and 5% higher F1-scores for rSS and long-range contact prediction, respectively. rMSA is available at https://zhanggroup.org/rMSA/ and https://github.com/pylelab/rMSA.

Molecular modeling and simulation play important roles in biomedical research as they provide molecular-level insight into the underlying mechanisms of biological functions that are difficult to elucidate only with experiments. CHARMM-GUI (https://charmm-gui.org) is a web-based cyberinfrastructure that is widely used to generate various molecular simulation system and input files and thus facilitates and standardizes the usage of common and advanced simulation techniques. In particular, PDB Manipulator provides various chemical modification options as the starting point for most input generation modules in CHARMM-GUI. Here, we discuss recent additions to PDB Manipulator, such as non-standard amino acids/RNA substitutions, ubiquitylation and SUMOylation, Lys/Arg post-translational modifications, lipidation, peptide stapling, and improved parameterization options of small molecules. These additional features are expected to make complex PDB modifications easy for biomolecular modeling and simulation.

ModelCIF (github.com/ihmwg/ModelCIF) is a data information framework developed for and by computational structural biologists to enable delivery of Findable, Accessible, Interoperable, and Reusable (FAIR) data to users worldwide. ModelCIF describes the specific set of attributes and metadata associated with macromolecular structures modeled by solely computational methods and provides an extensible data representation for deposition, archiving, and public dissemination of predicted three-dimensional (3D) models of macromolecules. It is an extension of the Protein Data Bank Exchange / macromolecular Crystallographic Information Framework (PDBx/mmCIF), which is the global data standard for representing experimentally-determined 3D structures of macromolecules and associated metadata. The PDBx/mmCIF framework and its extensions (e.g., ModelCIF) are managed by the Worldwide Protein Data Bank partnership (wwPDB, wwpdb.org) in collaboration with relevant community stakeholders such as the wwPDB ModelCIF Working Group (wwpdb.org/task/modelcif). This semantically rich and extensible data framework for representing computed structure models (CSMs) accelerates the pace of scientific discovery. Herein, we describe the architecture, contents, and governance of ModelCIF, and tools and processes for maintaining and extending the data standard. Community tools and software libraries that support ModelCIF are also described.

Ribosome profiling (Ribo-Seq) captures a "snapshot" of ribosomes' locations at the entire transcriptome of a cell at sub-codon resolution providing insights into gene expression and enabling the discovery of novel translated regions. RiboGalaxy (https://ribogalaxy.genomicsdatascience.ie/), a Galaxy-based platform for processing Ribo-Seq data is a RiboSeq.Org (https://riboseq.org/) resource. RiboSeq.Org is an online gateway to a set of integrated tools for the processing and analysis of Ribo-Seq data. In this RiboGalaxy update we introduce changes to both the tools available on RiboGalaxy and to how the resource is managed on the backend. For example, in order to improve interoperability between Riboseq.Org resources, we added tools that link RiboGalaxy outputs with Trips-Viz and GWIPS-viz browsers for downstream analysis and visualisation. RiboGalaxy's backend now utilises Ansible configuration management which enhances its stability and jobs are executed within Singularity containers and are managed by Slurm, strengthening reproducibility and performance respectively.

The knowledge of protein-protein interaction sites (PPIs) is crucial for protein functional annotation. Here we address the problem focusing on the prediction of putative PPIs considering as input protein sequences. The issue is important given the huge volume of protein sequences compared to experimental and/or computed structures. Taking advantage of protein language models, recently developed, and Deep Neural networks, here we describe ISPRED-SEQ, which overpasses state-of-the-art predictors addressing the same problem. ISPRED-SEQ is freely available for testing at https://ispredws.biocomp.unibo.it.

The Enzyme Function Initiative (EFI) provides a web resource with "genomic enzymology" web tools to leverage the protein (UniProt) and genome (European Nucleotide Archive; ENA; https://www.ebi.ac.uk/ena/) databases to assist the assignment of in vitro enzymatic activities and in vivo metabolic functions to uncharacterized enzymes (https://efi.igb.illinois.edu/). The tools enable (1) exploration of sequence-function space in enzyme families using sequence similarity networks (SSNs; EFI-EST), (2) easy access to genome context for bacterial, archaeal, and fungal proteins in the SSN clusters so that isofunctional families can be identified and their functions inferred from genome context (EFI-GNT); and (3) determination of the abundance of SSN clusters in NIH Human Metagenome Project metagenomes using chemically guided functional profiling (EFI-CGFP). We describe enhancements that enable SSNs to be generated from taxonomy categories, allowing higher resolution analyses of sequence-function space; we provide examples of the generation of taxonomy category-specific SSNs.

Metal-binding proteins are essential for the vital activities and engage in their roles by acting in concert with metal cations. MbPA (The Metal-binding Protein Atlas) is the most comprehensive resource up to now dedicated to curating metal-binding proteins. Currently, it contains 106,373 entries and 440,187 sites related to 54 metals and 8169 species. Users can view all metal-binding proteins and species-specific proteins in MbPA. There are also metal-proteomics data that quantitatively describes protein expression in different tissues and organs. By analyzing the data of the amino acid residues at the metal-binding site, it is found that about 80% of the metal ions tend to bind to cysteine, aspartic acid, glutamic acid, and histidine. Moreover, we use Diversity Measure to confirm that the diversity of metal-binding is specific in different area of periodic table, and further elucidate the binding modes of 19 transition metals on 20 amino acids. In addition, MbPA also embraces 6855 potential pathogenic mutations related to metalloprotein. The resource is freely available at http://bioinfor.imu.edu.cn/mbpa.

The sequence-dependent statistical mechanics of double-stranded nucleic acid, or dsNA, is believed to be essential in its biological functions. In turn, the equilibrium statistical mechanics behaviour of dsNA depends strongly both on sequence-dependent perturbations in its ground state shape away from an idealised, uniform, double-helical configuration, and on its fluctuations as governed by its sequence-dependent stiffness. We here describe the cgNA+web browser-based interactive tool for visualising the sequence-dependent ground states of dsNA fragments of arbitrary sequences, as predicted by the underlying cgNA+ coarse-grain model. Parameter sets are provided to model dsDNA, including the possibility of epigenetically modified CpG dinucleotide steps, dsRNA, and DNA:RNA Hybrid double helical fragments. The cgNA+web interface is specifically designed to compare ground state shapes of different sequences of the same dsNA, or analogous sequences of different dsNAs. The cgNA+web server is freely available at cgDNAweb.epfl.ch without any login requirement.

Protein structural domains have been less studied than full-length proteins in terms of ontology annotations. The dcGO database has filled this gap by providing mappings from protein domains to ontologies. The dcGO update in 2023 extends annotations for protein domains of multiple definitions (SCOP, Pfam, and InterPro) with commonly used ontologies that are categorised into functions, phenotypes, diseases, drugs, pathways, regulators, and hallmarks. This update adds new dimensions to the utility of both ontology and protein domain resources. A newly designed website at http://www.protdomainonto.pro/dcGO offers a more centralised and user-friendly way to access the dcGO database, with enhanced faceted search returning term- and domain-specific information pages. Users can navigate both ontology terms and annotated domains through improved ontology hierarchy browsing. A newly added facility enables domain-based ontology enrichment analysis.

The structural investigation of intrinsically disordered proteins (IDPs) requires ensemble models describing the diversity of the conformational states of the molecule. Due to their probabilistic nature, there is a need for new paradigms that understand and treat IDPs from a purely statistical point of view, considering their conformational ensembles as well-defined probability distributions. In this work, we define a conformational ensemble as an ordered set of probability distributions and provide a suitable metric to detect differences between two given ensembles at the residue level, both locally and globally. The underlying geometry of the conformational space is properly integrated, one ensemble being characterized by a set of probability distributions supported on the three-dimensional Euclidean space (for global-scale comparisons) and on the two-dimensional flat torus (for local-scale comparisons). The inherent uncertainty of the data is also taken into account to provide finer estimations of the differences between ensembles. Additionally, an overall distance between ensembles is defined from the differences at the residue level. We illustrate the potential of the approach with several examples of applications for the comparison of conformational ensembles: (i) produced from molecular dynamics (MD) simulations using different force fields, and (ii) before and after refinement with experimental data. We also show the usefulness of the method to assess the convergence of MD simulations, and discuss other potential applications such as in machine-learning-based approaches. The numerical tool has been implemented in Python through easy-to-use Jupyter Notebooks available at https://gitlab.laas.fr/moma/WASCO.

Operons are groups of consecutive genes that transcribe together under the regulation of a common promoter. They influence protein regulation and various physiological pathways, making their accurate detection desirable. The detection of operons through experimental means is a laborious and financially intensive process. Therefore, human experts predict potential operons utilizing their prior knowledge of the genetic organization and functional correlation of the genes. However, with the rise in the number of completely sequenced genomes, the development of automated algorithms, tools, and web servers is highly preferred over manual detection for operon prediction. Currently available state-of-the-art algorithms use a deep learning-based model to predict if the adjacent genes belong to the same operons in the given genome. However, these require an understanding of programming knowledge and computational skills, making them not-very user friendly. In this study, we developed a user-friendly web service, Operon Finder, for on-the-fly prediction of operons using the deep learning method. The interface provides a facility for genome search, operon live-filtering, viewing operonic DNA sequences, downloading predicted results, and links for data retrieval from the NCBI (National Center for Biotechnology Information) database. The web server is available at https://www.iitg.ac.in/spkanaujia/operonfinder.html. The experimental methods and the implementation details are publicly available at https://github.com/SPKlab/Operon-Finder.

The remarkable recent advances in protein structure prediction have enabled computational modeling of protein structures with considerably higher accuracy than ever before. While state-of-the-art structure prediction methods provide self-assessment confidence scores of their own predictions, an independent and open-access system for protein scoring is still needed that can be applied to a broad range of predictive modeling scenarios. Here, we present iQDeep, an integrated and highly customizable web server for protein scoring, freely available at http://fusion.cs.vt.edu/iQDeep. The underlying method of iQDeep employs multiscale deep residual neural networks (ResNets) to perform residue-level error classifications, and then probabilistically combines the error classifications for protein scoring. By adjusting the error resolutions, our method can reliably estimate the standard- or high-accuracy variants of the Global Distance Test metric for versatile protein scoring. The performance of the method has been extensively tested and compared against the state-of-the-art approaches in multiple rounds of Critical Assessment of Techniques for Protein Structure Prediction (CASP) experiments including benchmark assessment in CASP12 and CASP13 as well as blind evaluation in CASP14. The iQDeep web server offers a number of convenient features, including (i) the choice of individual and batch processing modes; (ii) an interactive and privacy-preserving web interface for automated job submission, tracking, and results retrieval; (iii) web-based quantitative and visual analyses of the results including overall estimated score and its residue-wise breakdown along with agreements between various sequence- and structural-level features; (iv) extensive help information on job submission and results interpretation via web-based tutorial and help tooltips.

Multiple sequence alignments (MSAs) are the workhorse of molecular evolution and structural biology research. From MSAs, the amino acids that are tolerated at each site during protein evolution can be inferred. However, little is known regarding the repertoire of tolerated amino acids in proteins when only a few or no sequence homologs are available, such as orphan and de novo designed proteins. Here we present EvoRator2, a deep-learning algorithm trained on over 15,000 protein structures that can predict which amino acids are tolerated at any given site, based exclusively on protein structural information mined from atomic coordinate files. We show that EvoRator2 obtained satisfying results for the prediction of position-weighted scoring matrices (PSSM). We further show that EvoRator2 obtained near state-of-the-art performance on proteins with high quality structures in predicting the effect of mutations in deep mutation scanning (DMS) experiments and that for certain DMS targets, EvoRator2 outperformed state-of-the-art methods. We also show that by combining EvoRator2's predictions with those obtained by a state-of-the-art deep-learning method that accounts for the information in the MSA, the prediction of the effect of mutation in DMS experiments was improved in terms of both accuracy and stability. EvoRator2 is designed to predict which amino-acid substitutions are tolerated in such proteins without many homologous sequences, including orphan or de novo designed proteins. We implemented our approach in the EvoRator web server (https://evorator.tau.ac.il).

Dimensionality reduction is a hot topic in machine learning that can help researchers find key features from complex medical or biological data, which is crucial for biological sequence research, drug development, etc. However, when applied to specific datasets, different dimensionality reduction methods generate different results, which produces instability and makes tuning the parameters a time-consuming task. Exploring high quality features, genes, or attributes from complex data is an important task and challenge. To ensure the efficiency, robustness, and accuracy of experiments, in this work, we developed a dimensionality reduction tool MRMD3.0 based on the ensemble strategy of link analysis. It is mainly divided into two steps: first, the ensemble method is used to integrate different feature ranking algorithms to calculate feature importance, and then the forward feature search strategy combined with cross-validation is used to explore the proper feature combination. Compared with the previously developed version, MRMD3.0 has added more link-based ensemble algorithms, including PageRank, HITS, LeaderRank, and TrustRank. At the same time, more feature ranking algorithms have been added, and their effect and calculation speed have been greatly improved. In addition, the newest version provides an interface used by each feature ranking method and five kinds of charts to help users analyze features. Finally, we also provide an online webserver to help researchers analyze the data. Availability and implementation Webserver: http://lab.malab.cn/soft/MRMDv3/home.html. GitHub: https://github.com/heshida01/MRMD3.0.

Ligand binding sites provide essential information for uncovering protein functions and structure-based drug discovery. To facilitate cavity detection and property analysis process, we developed a comprehensive web server, CavityPlus in 2018. CavityPlus applies the CAVITY program to detect potential binding sites in a given protein structure. The CavPharmer, CorrSite, and CovCys tools can then be applied to generate receptor-based pharmacophore models, identify potential allosteric sites, or detect druggable cysteine residues for covalent drug design. While CavityPlus has been widely used, the constantly evolving knowledge and methods make it necessary to improve and extend its functions. This study presents a new version of CavityPlus, CavityPlus 2022 through a series of upgrades. We upgraded the CAVITY tool to greatly speed up cavity detection calculation. We optimized the CavPharmer tool for fast speed and more accurate results. We integrated the newly developed CorrSite2.0 into the CavityPlus 2022 web server for its improved performance of allosteric site prediction. We also added a new CavityMatch module for drug repurposing and protein function studies by searching similar cavities to a given cavity from pre-constructed cavity databases. The new version of CavityPlus is freely available at http://pkumdl.cn:8000/cavityplus/.

In 2019, we released Missense3D which identifies stereochemical features that are disrupted by a missense variant, such as introducing a buried charge. Missense3D analyses the effect of a missense variant on a single structure and thus may fail to identify as damaging surface variants disrupting a protein interface i.e., a protein-protein interaction (PPI) site. Here we present Missense3D-PPI designed to predict missense variants at PPI interfaces. Our development dataset comprised of 1,279 missense variants (pathogenic n = 733, benign n = 546) in 434 proteins and 545 experimental structures of PPI complexes. Benchmarking of Missense3D-PPI was performed after dividing the dataset in training (320 benign and 320 pathogenic variants) and testing (226 benign and 413 pathogenic). Structural features affecting PPI, such as disruption of interchain bonds and introduction of unbalanced charged interface residues, were analysed to assess the impact of the variant at PPI. The performance of Missense3D-PPI was superior to that of Missense3D: sensitivity 44 % versus 8% and accuracy 58% versus 40%, p = 4.23 × 10-16. However, the specificity of Missense3D-PPI was lower compared to Missense3D (84% versus 98%). On our dataset, Missense3D-PPI's accuracy was superior to BeAtMuSiC (p = 3.4 × 10-5), mCSM-PPI2 (p = 1.5 × 10-12) and MutaBind2 (p = 0.0025). Missense3D-PPI represents a valuable tool for predicting the structural effect of missense variants on biological protein networks and is available at the Missense3D web portal (http://missense3d.bc.ic.ac.uk).

The human interactome is composed of around half a million interactions according to recent estimations and it is only for a small fraction of those that three-dimensional structural information is available. Indeed, the structural coverage of the human interactome is very low and given the complexity and time-consuming requirements of solving protein structures this problem will remain for the foreseeable future. Structural models, or predictions, of protein complexes can provide valuable information when the experimentally determined 3D structures are not available. Here we present CM2D3, a relational database containing structural models of the whole human interactome derived both from comparative modeling and data-driven docking. Starting from a consensus interactome derived from integrating several interactomics databases, a strategy was devised to derive structural models by computational means. Currently, CM2D3 includes 33338 structural models of which 5121 derived from comparative modeling and the remaining from docking. Of the latter, the structures of 14554 complexes were derived from monomers modeled by M4T while the rest were modeled with structures as predicted by AlphaFold2. Lastly, CM2D3 complements existing resources by focusing on models derived from both free-docking, as opposed to template-based docking, and hence expanding the available structural information on protein complexes to the scientific community. Database URL:http://www.bioinsilico.org/CM2D3.

Dietary components and bioactive molecules present in functional foods and nutraceuticals provide various beneficial effects including modulation of host gut microbiome. These metabolites along with orally administered drugs can be potentially bio-transformed by gut microbiome, which can alter their bioavailability and intended biological or pharmacological activity resulting in individual or population-specific variation in drug and dietary responses. Experimental determination of microbiome-mediated metabolism of orally ingested molecules is difficult due to the enormous diversity and complexity of the gut microbiome. To address this problem, we developed "GutBug", a web-based resource that predicts all possible bacterial metabolic enzymes that can potentially biotransform xenobiotics and biotic molecules using a combination of machine learning, neural networks and chemoinformatic methods. Using 3,457 enzyme substrates for training and a curated database of 363,872 enzymes from ∼700 gut bacterial strains, GutBug can predict complete EC number of the bacterial enzymes involved in a biotransformation reaction of the given molecule along with the reaction centres with accuracies between 0.78 and 0.97 across different reaction classes. Validation of GutBug's performance using 27 molecules known to be biotransformed by human gut bacteria, including complex polysaccharides, flavonoids, and oral drugs further attests to GutBug's accuracy and utility. Thus, GutBug enhances our understanding of various metabolite-gut bacterial interactions and their resultant effects on the human host health across populations, which will find enormous applications in diet design and intervention, identification and administration of new prebiotics, development of nutraceutical products, and improvements in drug designing. GutBug is available at https://metabiosys.iiserb.ac.in/gutbug.

Over 2500 Salmonella species (alternatively, serovars) encompassing different combinations of O-, H1- and H2-antigens are present in nature and cause millions of deaths worldwide every year. Since conventional serotyping is time-consuming, a user-friendly Salmonellaspecies serotyping (SSP) web tool (https://project.iith.ac.in/SSP/) is developed here to predict the serotypes using Salmonella protein(s) or whole proteome sequences. Prior to SSP implementation, a detailed analysis of protein sequences involved in O-antigen biosynthesis and H-antigen formation is carried out to assess their serotype specificity. Intriguingly, the results indicate that the initializing transferases WbaP, WecA and GNE can efficiently distinguish the O-antigens, which have Gal, GlcNAc and GalNAc as initial sugars respectively. Rigorous analysis shows that Wzx and Wzy are sufficient to distinguish the O-types. Exceptionally, some situations warrant additional proteins. Thus, 150 additional transferases, RfbE for O2, O9 and O9,46 types, Orf17.4 for O3,10 and O1,3,19 types, WecB, WbbE and WbbF for O54 and, Wzm and Wzt for O67 are utilized in serotyping. An in-depth analysis of 302 reference datasets representing 56 H1- and 20 H2-types leads to the identification and utilization of 61 unique sequence patterns of FliC and FljB in H-typing. A test dataset of 2136 whole proteome sequences covering 740 Salmonella serovars, including 13 new species are successfully predicted with 99.72% accuracy. Prior to this, all the O-, H1- and H2-antigens are predicted accurately when tested independently. Indeed, SSP also identifies wrongly annotated Salmonella species; hence, it can easily identify new species that emerge with any combination of O-, H1- and H2-antigens. Thus, SSP can act as a valuable tool in the surveillance of Salmonella species.

Biofilms are one of the leading causes of antibiotic resistance. It acts as a physical barrier against the human immune system and drugs. The use of anti-biofilm agents helps in tackling the menace of antibiotic resistance. The identification of efficient anti-biofilm chemicals remains a challenge. Therefore, in this study, we developed 'anti-Biofilm', a machine learning technique (MLT) based predictive algorithm for identifying and analyzing the biofilm inhibition of small molecules. The algorithm is developed using experimentally validated anti-biofilm compounds with half maximal inhibitory concentration (IC50) values extracted from aBiofilm resource. Out of the five MLTs, the Support Vector Machine performed best with Pearson's correlation coefficient of 0.75 on the training/testing data set. The robustness of the developed model was further checked using an independent validation dataset. While analyzing the chemical diversity of the anti-biofilm compounds, we observed that they occupy diverse chemical spaces with parent molecules like furanone, urea, phenolic acids, quinolines, and many more. Use of diverse chemicals as input further signifies the robustness of our predictive models. The three best-performing machine learning models were implemented as a user-friendly 'anti-Biofilm' web server (https://bioinfo.imtech.res.in/manojk/antibiofilm/) with different other modules which make 'anti-Biofilm' a comprehensive platform. Therefore, we hope that our initiative will be helpful for the scientific community engaged in identifying effective anti-biofilm agents to target the problem of antimicrobial resistance.

In the ten years since the discovery of the first anti-CRISPR (Acr) proteins, the number of validated Acrs has expanded rapidly, as has our understanding of the diverse mechanisms they employ to suppress natural CRISPR-Cas immunity. Many, though not all, function via direct, specific interaction with Cas protein effectors. The abilities of Acr proteins to modulate the activities and properties of CRISPR-Cas effectors have been exploited for an ever-increasing spectrum of biotechnological uses, most of which involve the establishment of control over genome editing systems. This control can be used to minimize off-target editing, restrict editing based on spatial, temporal, or conditional cues, limit the spread of gene drive systems, and select for genome-edited bacteriophages. Anti-CRISPRs have also been developed to overcome bacterial immunity, facilitate viral vector production, control synthetic gene circuits, and other purposes. The impressive and ever-growing diversity of Acr inhibitory mechanisms will continue to allow the tailored applications of Acrs.

SARS-CoV-2 virus spike (S) protein is an envelope protein responsible for binding to the ACE2 receptor, driving subsequent entry into host cells. The existence of multiple disulfide bonds in the S protein makes it potentially susceptible to reductive cleavage. Using a tri-part split luciferase-based binding assay, we evaluated the impacts of chemical reduction on S proteins from different virus variants and found that those from the Omicron family are highly vulnerable to reduction. Through manipulation of different Omicron mutations, we found that alterations in the receptor binding module (RBM) are the major determinants of this vulnerability. Specifically we discovered that Omicron mutations facilitate the cleavage of C480-C488 and C379-C432 disulfides, which consequently impairs binding activity and protein stability. The vulnerability of Omicron S proteins suggests a mechanism that can be harnessed to treat specific SARS-CoV-2 strains.

Vaccines are among the greatest tools for prevention and control of disease. They have eliminated smallpox from the planet, decreased morbidity and mortality for major infectious diseases like polio, measles, mumps, and rubella, significantly blunted the impact of the COVID-19 pandemic, and prevented viral induced cancers such as cervical cancer caused by human papillomavirus. Recent technological advances, in genomics, structural biology, and human immunology have transformed vaccine development, enabling new technologies such as mRNA vaccines to greatly accelerate development of new and improved vaccines. In this review, we briefly highlight the history of vaccine development, and provide examples of where advances in genomics and structural biology, paved the way for development of vaccines for bacterial and viral diseases.

Adverse pregnancy outcomes including maternal mortality, stillbirth, preterm birth, intrauterine growth restriction cause millions of deaths each year. More effective interventions are urgently needed. Maternal immunization could be one such intervention protecting the mother and newborn from infection through its pathogen-specific effects. However, many adverse pregnancy outcomes are not directly linked to the infectious pathogens targeted by existing maternal vaccines but rather are linked to pathological inflammation unfolding during pregnancy. The underlying pathogenesis driving such unfavourable outcomes have only partially been elucidated but appear to relate to altered immune regulation by innate as well as adaptive immune responses, ultimately leading to aberrant maternal immune activation. Maternal immunization, like all immunization, impacts the immune system beyond pathogen-specific immunity. This raises the possibility that maternal vaccination could potentially be utilised as a pathogen-agnostic immune modulatory intervention to redirect abnormal immune trajectories towards a more favourable phenotype providing pregnancy protection. In this review we describe the epidemiological evidence surrounding this hypothesis, along with the mechanistic plausibility and present a possible path forward to accelerate addressing the urgent need of adverse pregnancy outcomes.

SNARE is the essential mediator of membrane fusion that highly relies on the molecular structure of SNAREs. For instance, the protein syntaxin-1 involved in neuronal SNAREs, has a single transmembrane domain (sTMD) leading to fast fusion, while the syntaxin 17 has a V-shape double TMDs (dTMDs), taking part in the autophagosome maturation. However, it is not clear how the TMD structure influences the fusion process. Here, we demonstrate that the dTMDs significantly reduce fusion rate compared with the sTMD by using an in vitro reconstitution system. Through theoretical analysis, we reveal that the V-shape dTMDs can significantly increase protein-lipid mismatch, thereby raising the energy barrier of the fusion, and that increasing the number of SNAREs can reduce the energy barrier or protein-lipid mismatch. This study provides a physicochemical mechanistic understanding of SNARE-regulated membrane fusion.

Macromolecular interactions regulate all aspects of biology. The identification of interacting partners and complexes is important for understanding cellular processes, host-pathogen conflicts, and organismal development. Multiple methods exist to label and enrich interacting proteins in living cells. Notably, the soybean ascorbate peroxidase, APEX2, rapidly biotinylates adjacent biomolecules in the presence of biotin-phenol and hydrogen peroxide. However, during initial experiments with this system, we found that APEX2 exhibits a cytoplasmic-biased localization and is sensitive to the nuclear export inhibitor leptomycin B (LMB). This led us to identify a putative nuclear export signal (NES) at the carboxy-terminus of APEX2 (NESAPEX2), structurally adjacent to the conserved heme binding site. This putative NES is functional as evidenced by cytoplasmic localization and LMB sensitivity of a mCherry-NESAPEX2 chimeric construct. Single amino acid substitutions of multiple hydrophobic residues within NESAPEX2 eliminate cytoplasm-biased localization of both mCherry-NESAPEX2 as well as full-length APEX2. However, all but one of these NES substitutions also compromises peroxide-dependent labeling. This unique separation-of-function mutant, APEX2-L242A, is termed APEX3. Localization and functionality of APEX3 are confirmed by fusion to the nucleocytoplasmic shuttling transcriptional factor, RELA. APEX3 is therefore an optimized tool for unbiased proximity labeling of cellular proteins and interacting factors..

The molecular basis for septin filament assembly has begun to emerge over recent years. These filaments are essential for many septin functions which depend on their association with biological membranes or components of the cytoskeleton. Much less is known about how septins specifically interact with their binding partners. Here we describe the essential role played by the C-terminal domains in both septin polymerization and their association with the BD3 motif of the Borg family of Cdc42 effector proteins. We provide a detailed description, at the molecular level, of a previously reported interaction between BD3 and the NC-interface between SEPT6 and SEPT7. Upon ternary complex formation, the heterodimeric coiled coil formed by the C-terminal domains of the septins becomes stabilized and filament formation is promoted under conditions of ionic strength/protein concentration which are not normally permissible, likely by favouring hexamers over smaller oligomeric states. This demonstrates that binding partners, such as Borg's, have the potential to control filament assembly/disassembly in vivo in a way which can be emulated in vitro by altering the ionic strength. Experimentally validated models indicate that the BD3 peptide lies antiparallel to the coiled coil and is stabilized by a mixture of polar and apolar contacts. At its center, an LGPS motif, common to all human Borg sequences, interacts with charged residues from both helices of the coiled coil (K368 from SEPT7 and the conserved E354 from SEPT6) suggesting a universal mechanism which governs Borg-septin interactions.

In transcription-coupled repair (TCR), transcribing RNA polymerase II (RNAPII) stalls at a DNA lesion and recruits TCR proteins to the damaged site. However, the mechanism by which RNAPII recognizes a DNA lesion in the nucleosome remains enigmatic. In the present study, we inserted an apurinic/apyrimidinic DNA lesion analogue, tetrahydrofuran (THF), in the nucleosomal DNA, where RNAPII stalls at the SHL(-4), SHL(-3.5), and SHL(-3) positions, and determined the structures of these complexes by cryo-electron microscopy. In the RNAPII-nucleosome complex stalled at SHL(-3.5), the nucleosome orientation relative to RNAPII is quite different from those in the SHL(-4) and SHL(-3) complexes, which have nucleosome orientations similar to naturally paused RNAPII-nucleosome complexes. Furthermore, we found that an essential TCR protein, Rad26 (CSB), enhances the RNAPII processivity, and consequently augments the DNA damage recognition efficiency of RNAPII in the nucleosome. The cryo-EM structure of the Rad26-RNAPII-nucleosome complex revealed that Rad26 binds to the stalled RNAPII through a novel interface, which is completely different from those previously reported. These structures may provide important information to understand the mechanism by which RNAPII recognizes the nucleosomal DNA lesion and recruits TCR proteins to the stalled RNAPII on the nucleosome.

Mitochondrial protein import is critical for organelle biogenesis, bioenergetic function, and health. The mechanism of which is poorly understood, particularly of the mammalian system. To address this problem we have established an assay to quantitatively monitor mitochondrial import inside mammalian cells. The reporter is based on a split luciferase, whereby the large fragment is segregated in the mitochondrial matrix and the small complementary fragment is fused to the C-terminus of a purified recombinant precursor protein destined for import. Following import the complementary fragments combine to form an active luciferase-providing a sensitive, accurate and continuous measure of protein import. This advance allows detailed mechanistic examination of the transport process in live cells, including the analysis of import breakdown associated with disease, and high-throughput drug screening. Furthermore, the set-up has the potential to be adapted for the analysis of alternative protein transport systems within different cell types, and multicellular model organisms.

Small heat shock proteins (sHSPs) are essential ATP-independent chaperones that protect the cellular proteome. These proteins assemble into polydisperse oligomeric structures, the composition of which dramatically affects their chaperone activity. The biomolecular consequences of variations in sHSP ratios, especially inside living cells, remain elusive. Here, we study the consequences of altering the relative expression levels of HspB2 and HspB3 in HEK293T cells. These chaperones are partners in a hetero-oligomeric complex, and genetic mutations that abolish their mutual interaction are associated with myopathic disorders. HspB2 displays three distinct phenotypes when co-expressed with HspB3 at varying ratios. Expression of HspB2 alone leads to formation of liquid nuclear condensates, while shifting the stoichiometry towards HspB3 resulted in the formation of large solid-like aggregates. Only cells co-expressing HspB2 with a limited amount of HspB3 formed fully soluble complexes that were distributed homogeneously throughout the nucleus. Strikingly, both condensates and aggregates were reversible, as shifting the HspB2:HspB3 balance in situ resulted in dissolution of these structures. To uncover the molecular composition of HspB2 condensates and aggregates, we used APEX-mediated proximity labelling. Most proteins interact transiently with the condensates and were neither enriched nor depleted in these cells. In contrast, we found that HspB2:HspB3 aggregates sequestered several disordered proteins and autophagy factors, suggesting that the cell is actively attempting to clear these aggregates. This study presents a striking example of how changes in the relative expression levels of interacting proteins affects their phase behavior. Our approach could be applied to study the role of protein stoichiometry and the influence of client binding on phase behavior in other biomolecular condensates and aggregates.

Identifying the interactions between proteins and ligands is significant for drug discovery and design. Considering the diverse binding patterns of ligands, the ligand-specific methods are trained per ligand to predict binding residues. However, most of the existing ligand-specific methods ignore shared binding preferences among various ligands and generally only cover a limited number of ligands with a sufficient number of known binding proteins. In this study, we propose a relation-aware framework LigBind with graph-level pre-training to enhance the ligand-specific binding residue predictions for 1159 ligands, which can effectively cover the ligands with a few known binding proteins. LigBind first pre-trains a graph neural network-based feature extractor for ligand-residue pairs and relation-aware classifiers for similar ligands. Then, LigBind is fine-tuned with ligand-specific binding data, where a domain adaptive neural network is designed to automatically leverage the diversity and similarity of various ligand-binding patterns for accurate binding residue prediction. We construct ligand-specific benchmark datasets of 1159 ligands and 16 unseen ligands, which are used to evaluate the effectiveness of LigBind. The results demonstrate the LigBind's efficacy on large-scale ligand-specific benchmark datasets, and it generalizes well to unseen ligands. LigBind also enables accurate identification of the ligand-binding residues in the main protease, papain-like protease and the RNA-dependent RNA polymerase of SARS-CoV-2. The web server and source codes of LigBind are available at http://www.csbio.sjtu.edu.cn/bioinf/LigBind/ and https://github.com/YYingXia/LigBind/ for academic use.

Transcription factors (TF) recognize specific motifs in the genome that are typically 6-12 bp long to regulate various aspects of the cellular machinery. Presence of binding motifs and favorable genome accessibility are key drivers for a consistent TF-DNA interaction. Although these pre-requisites may occur thousands of times in the genome, there seems to be a high degree of selectivity for the sites that are actually bound. Here, we present a deep-learning framework that identifies and characterizes the upstream and downstream genetic elements to the binding motif, for their role in enforcing the mentioned selectivity. The proposed framework is based on an interpretable recurrent neural network architecture that enables for the relative analysis of sequence context features. We apply the framework to model twenty-six transcription factors and score the TF-DNA binding at a base-pair resolution. We find significant differences in activations of DNA context features for bound and unbound sequences. In addition to standardized evaluation protocols, we offer outstanding interpretability that enables us to identify and annotate DNA sequence with possible elements that modulate TF-DNA binding. Also, differences in data processing have a huge influence on the overall model performance. Overall, the proposed framework allows for novel insights on the non-coding genetic elements and their role in facilitating a stable TF-DNA interaction.

Sensing extracellular glucose, budding yeast switches from aerobic glycolysis to oxidative phosphorylation to adapt to environmental changes. During the conversion of metabolic mode, mitochondrial function and morphology change significantly. Mitochondria are the main supply factories of energy for various life activities in cells. However, the research on the signal pathways from glucose sensing to changes in mitochondrial function and morphology is still scarce and worthy of further exploration. In this study, we found that in addition to the known involvement of molecular chaperone Hsp82 in stress response during the conversion of metabolic mode, the phosphorylation status of Hsp82 at S485 residue regulates mitochondrial function and morphology to maintain mitochondrial homeostasis. The Hsp82S485A mutant that mimics dephosphorylation at S485 residue showed abnormal growth phenotypes related to mitochondrial defects, such as the petite phenotype, slow growth rates, and inability to use non-fermentable carbon sources. Further exploring the causes of growth defects, we found that the Hsp82S485A mutant caused mitochondrial dysfunction, including a decrease in cellular oxygen consumption rate, defects in mitochondrial electron transport chain, decreased mitochondrial membrane potential and complete loss of mtDNA. Furthermore, the Hsp82S485A mutant displayed fragmented or globular mitochondria, which may be responsible for its mitochondrial dysfunction. Our findings suggested that the phosphorylation status of Hsp82 at S485 residue might regulate mitochondrial function and morphology by affecting the stability of mitochondrial fission and fusion-related proteins. Thus, Hsp82 might be a key molecule in the signal pathway from glucose sensing to changes in mitochondrial function and morphology.

Human papilloma virus (HPV) infections are associated with almost all cervical cancers and to a lower extend also with anogenital or oropharyngeal cancers. HPV proteins expressed in HPV-associated tumors are attractive antigens for cancer vaccination strategies as self-tolerance, which is associated with most endogenous tumor-associated antigens, does not need to be overcome. In this study, we generated a live attenuated cancer vaccine based on the chimeric vesicular stomatitis virus VSV-GP, which has previously proven to be a potent vaccine vector and oncolytic virus. Genes at an earlier position in the genome more to the 3' end are expressed stronger compared to genes located further downstream. By inserting an HPV16-derived antigen cassette consisting of E2, E6 and E7 into VSV-GP either at first (HPVp1) or fifth (HPVp5) position in VSV-GP's genome we aimed to analyze the effect of vaccine antigen position and consequently expression level on viral fitness, immunogenicity, and anti-tumoral efficacy in a syngeneic mouse tumor model. HPVp1 expressed higher amounts of HPV antigens compared to HPVp5 in vitro but had a slightly delayed replication kinetic which overall translated into increased HPV-specific T cell responses upon vaccination of mice. Immunization with both vectors protected mice in prophylactic and in therapeutic TC-1 tumor models with HPVp1 being more effective in the prophylactic setting. Taken together, VSV-GP is a promising candidate as therapeutic HPV vaccine and first position of the vaccine antigen in a VSV-derived vector seems to be superior to fifth position.

Matrix metalloproteinases (MMPs) are key drivers of various diseases, including cancer. Development of probes and drugs capable of selectively inhibiting the individual members of the large MMP family remains a persistent challenge. The inhibitory N-terminal domain of tissue inhibitor of metalloproteinases-2 (N-TIMP2), a natural broad MMP inhibitor, can provide a scaffold for protein engineering to create more selective MMP inhibitors. Here, we pursued a unique approach harnessing both computational design and combinatorial screening to confer high binding specificity toward a target MMP in preference to an anti-target MMP. We designed a loop extension of N-TIMP2 to allow new interactions with the non-conserved MMP surface and generated an efficient focused library for yeast surface display, which was then screened for high binding to the target MMP-14 and low binding to anti-target MMP-3. Deep sequencing analysis identified the most promising variants, which were expressed, purified, and tested for selectivity of inhibition. Our best N-TIMP2 variant exhibited 29 pM binding affinity to MMP-14 and 2.4 µM affinity to MMP-3, revealing 7500-fold greater specificity than WT N-TIMP2. High-confidence structural models were obtained by including NGS data in the AlphaFold multiple sequence alignment. The modeling together with experimental mutagenesis validated our design predictions, demonstrating that the loop extension packs tightly against non-conserved residues on MMP-14 and clashes with MMP-3. This study demonstrates how introduction of loop extensions in a manner guided by target protein conservation data and loop design can offer an attractive strategy to achieve specificity in design of protein ligands.