Protein Sequence Similarity Research Articles

Molecular evolution analysis of MYB5 in Brassicaceae with specific focus on seed coat color of Brassica napus

BackgroundMYB transcription factors are splay a vital role in plant biology, with previous research highlighting the significant impact of the R2R3-MYB-like transcription factor MYB5 on seed mucilage biosynthesis, trichome branching, and seed coat development. However, there is a dearth of studies investigating its role in the regulation of proanthocyanidin (PA) biosynthesis.ResultsIn this study, a total of 51 MYB5 homologous genes were identified across 31 species belonging to the Brassicaceae family, with particular emphasis on Brassica napus for subsequent investigation. Through phylogenetic analysis, these genes were categorized into four distinct subclasses. Protein sequence similarity and identity analysis demonstrated a high degree of conservation of MYB5 among species within the Brassicaceae family. Additionally, the examination of selection pressure revealed that MYB5 predominantly underwent purifying selection during its evolutionary history, as indicated by the Ka/Ks values of all MYB5 homologous gene pairs being less than one. Notably, we observed a higher rate of non-synonymous mutations in orthologous genes compared to paralogous genes, and the Ka/Ks value displayed a stronger correlation with Ka. In B. napus, an examination of expression patterns in five tissues revealed that MYB5 exhibited particularly high expression in the black seed coat. The findings from the WGCNA demonstrated a robust correlation between MYB5 and BAN(ANR) associated with PA biosynthesis in the black seed coat, providing further evidence of their close association and co-expression. Furthermore, the results obtained from of the analysis of protein interaction networks offer supplementary support for the proposition that MYB5 possesses the capability to interact with transcriptional regulatory proteins, specifically TT8 and TT2, alongside catalytic enzymes implicated in the synthesis of PAs, thereby making a contribution to the biosynthesis of PAs. These findings imply a plausible and significant correlation between the nuique expression pattern of MYB5 and the pigmentation of rapeseed coats. Nevertheless, additional research endeavors are imperative to authenticate and substantiate these findings.ConclusionsThis study offers valuable insights into the genetic evolution of Brassicaceae plants, thereby serving as a significant reference for the genetic enhancement of Brassicaceae seed coat color.

GSLAlign: community detection and local PPI network alignment

High throughput protein–protein interaction (PPI) profiling and computational techniques have resulted in generating a large amount of PPI network data. The study of PPI networks helps in understanding the biological processes of the proteins. The comparative study of the PPI networks helps in identifying the conserved interactions across the species. This article presents a novel local PPI network aligner ‘GSLAlign’ that consists of two stages. It first detects the communities from the PPI networks by applying the GraphSAGE algorithm using gene expression data. In the second stage, the detected communities are aligned using a community aligner that is based on protein sequence similarity. The community detection algorithm produces more separable and biologically accurate communities as compared to previous community detection algorithms. Moreover, the proposed community alignment algorithm achieves 3–8% better results in terms of semantic similarity as compared to previous local aligners. The average connectivity and coverage of the proposed algorithm are also better than the existing aligners. Communicated by Ramaswamy H. Sarma

Purification of Hemocyanin‐Derived Phenoloxidase and Comparison of Its Properties With Polyphenoloxidase in the Shrimp Litopenaeus vannamei

Purification of Hemocyanin‐Derived Phenoloxidase and Comparison of Its Properties With Polyphenoloxidase in the Shrimp <i>Litopenaeus vannamei</i>

In Silico Analysis of Cross-Species Sequence Variability in Host Interferon Antiviral Pathway Proteins and SARS-CoV-2 Susceptibility

Background: Zoonotic transmission of severe acute respiratory coronavirus 2 (SARS-CoV-2) has been found to result in infections in more than 30 mammalian species. The SARS-CoV-2 spike protein binds to the host’s angiotensin converting enzyme 2 (ACE2) cell surface receptor to gain entry into the cell. ACE2 protein sequence conservation has therefore been evaluated across species, and species with amino acid substitutions in ACE2 were ranked low for susceptibility to SARS-CoV-2 infection. However, many of these species have become infected by the virus. Methods: This study investigated the conservation of 24 host protein targets, including the entry proteins ACE2 and transmembrane serine protease 2 (TMPRSS2); 21 proteins in the interferon-I (IFN-I) antiviral response pathway; and tethrin, a protein that suppresses new virion release from cells. Bioinformatics approaches including Sequence Alignment to Predict Across Species Susceptibility (SeqAPASS), Molecular Operating Environment (MOE), and iCn3D software were used to compare protein sequence similarity, conserved domains, and critical amino acids for host-viral protein-protein interactions. The types of bonding interactions were scored, and the results were compared with empirical data indicating which species have or have not become infected. Results: This pathway approach revealed that 1) 13 proteins were conserved, whereas five lacked data sufficient to determine specific critical amino acids; 2) variation in protein-protein interfaces is tolerated for many amino acid substitutions, and these substitutions follow taxonomic clades rather than correlating with empirically determined species infection status; and 3) four proteins (MDA5, NEMO, IRF3, and ISG15) contained potential domains or specific amino acids whose substitution may result in PPI disruption. Conclusion: This work provides evidence that certain substitutions in four IFN-I antiviral pathway proteins appear able to disrupt interactions and may be distinctive to resistant species, thus potentially aiding in determining species’ likelihood of transmitting SARS-CoV-2.

SARS-CoV-2 helicase might interfere with cellular nonsense-mediated RNA decay: insights from a bioinformatics study

BackgroundViruses employ diverse strategies to interfere with host defense mechanisms, including the production of proteins that mimic or resemble host proteins. This study aimed to analyze the similarities between SARS-CoV-2 and human proteins, investigate their impact on virus-host interactions, and elucidate underlying mechanisms.ResultsComparing the proteins of SARS-CoV-2 with human and mammalian proteins revealed sequence and structural similarities between viral helicase with human UPF1. The latter is a protein that is involved in nonsense-mediated RNA decay (NMD), an mRNA surveillance pathway which also acts as a cellular defense mechanism against viruses. Protein sequence similarities were also observed between viral nsp3 and human Poly ADP-ribose polymerase (PARP) family of proteins. Gene set enrichment analysis on transcriptomic data derived from SARS-CoV-2 positive samples illustrated the enrichment of genes belonging to the NMD pathway compared with control samples. Moreover, comparing transcriptomic data from SARS-CoV-2-infected samples with transcriptomic data derived from UPF1 knockdown cells demonstrated a significant overlap between datasets.ConclusionsThese findings suggest that helicase/UPF1 sequence and structural similarity might have the ability to interfere with the NMD pathway with pathogenic and immunological implications.

Humanization reveals pervasive incompatibility of yeast and human kinetochore components.

Kinetochores assemble on centromeres to drive chromosome segregation in eukaryotic cells. Humans and budding yeast share most of the structural subunits of the kinetochore, whereas protein sequences have diverged considerably. The conserved centromeric histone H3 variant, CenH3 (CENP-A in humans and Cse4 in budding yeast), marks the site for kinetochore assembly in most species. A previous effort to complement Cse4 in yeast with human CENP-A was unsuccessful; however, co-complementation with the human core nucleosome was not attempted. Previously, our lab successfully humanized the core nucleosome in yeast; however, this severely affected cellular growth. We hypothesized that yeast Cse4 is incompatible with humanized nucleosomes and that the kinetochore represented a limiting factor for efficient histone humanization. Thus, we argued that including the human CENP-A or a Cse4-CENP-A chimera might improve histone humanization and facilitate kinetochore function in humanized yeast. The opposite was true: CENP-A expression reduced histone humanization efficiency, was toxic to yeast, and disrupted cell cycle progression and kinetochore function in wild-type (WT) cells. Suppressors of CENP-A toxicity included gene deletions of subunits of 3 conserved chromatin remodeling complexes, highlighting their role in CenH3 chromatin positioning. Finally, we attempted to complement the subunits of the NDC80 kinetochore complex, individually and in combination, without success, in contrast to a previous study indicating complementation by the human NDC80/HEC1 gene. Our results suggest that limited protein sequence similarity between yeast and human components in this very complex structure leads to failure of complementation.

Developing similarity matrices for antibody-protein binding interactions.

The inventions of AlphaFold and RoseTTAFold are revolutionizing computational protein science due to their abilities to reliably predict protein structures. Their unprecedented successes are due to the parallel consideration of several types of information, one of which is protein sequence similarity information. Sequence homology has been studied for many decades and depends on similarity matrices to define how similar or different protein sequences are to one another. A natural extension of predicting protein structures is predicting the interactions between proteins, but similarity matrices for protein-protein interactions do not exist. This study conducted a mutational analysis of 384 non-redundant antibody-protein antigen complexes to calculate antibody-protein interaction similarity matrices. Every important residue in each antibody and each antigen was mutated to each of the other 19 commonly occurring amino acids and the percentage changes in interaction energies were calculated using three force fields: CHARMM, Amber, and Rosetta. The data were used to construct six interaction similarity matrices, one for antibodies and another for antigens using each force field. The matrices exhibited both commonalities, such as mutations of aromatic and charged residues being the most detrimental, and differences, such as Rosetta predicting mutations of serines to be better tolerated than either Amber or CHARMM. A comparison to nine previously published similarity matrices for protein sequences revealed that the new interaction matrices are more similar to one another than they are to any of the previous matrices. The created similarity matrices can be used in force field specific applications to help guide decisions regarding mutations in protein-protein binding interfaces.

PC_ali: A tool for improved multiple alignments and evolutionary inference based on a hybrid protein sequence and structure similarity score.

Evolutionary inference depends crucially on the quality of multiple sequence alignments (MSA), which is problematic for distantly related proteins. Since protein structure is more conserved than sequence, it seems natural to use structure alignments for distant homologs. However, structure alignments may not be suitable for inferring evolutionary relationships. Here we examined four protein similarity measures that depend on sequence and structure (fraction of aligned residues, sequence identity, fraction of superimposed residues and contact overlap), finding that they are intimately correlated but none of them provides a complete and unbiased picture of conservation in proteins. Therefore, we propose the new hybrid protein sequence and structure similarity score PC_sim based on their main Principal Component. The corresponding divergence measure PC_div shows the strongest correlation with divergences obtained from individual similarities, suggesting that it infers accurate evolutionary divergences. We developed the program PC_ali that constructs protein MSAs either de novo or modifying an input MSA, using a similarity matrix based on PC_sim. tructs a starting MSA based on the maximal cliques of the graph of these PAs and it refines it through progressive alignments along the tree reconstructed with PC_div. Compared with eight state-of-the-art multiple structure or sequence alignment tools, PC_ali achieve higher or equal aligned fraction and structural scores, sequence identity higher than structure aligners although lower than sequence aligners, highest score PC_sim and highest similarity with the MSAs produced by other tools and with the reference MSA Balibase. https://github.com/ugobas/PC_ali. Supplementary data are available at Bioinformatics online.

Anti-parasitic benzoxaboroles are ineffective against Theileria parva in vitro

East Coast Fever (ECF) is a disease affecting cattle in sub-Saharan Africa, caused by the tick-borne Apicomplexan pathogen Theileria parva. The disease is a major problem for cattle farmers in affected regions and there are few methods of control, including a complex infection and treatment vaccine, expensive chemotherapy, and the more widespread tick control through acaricides. New intervention strategies are, therefore, sorely needed. Benzoxaboroles are a versatile class of boron-heterocyclic compounds with demonstrable pharmacological activity against a diverse group of pathogens, including those related to T. parva. In this study, the in vitro efficacy of three benzoxaboroles against the intracellular schizont stage of T. parva was investigated using a flow cytometry approach. Of the benzoxaboroles tested, only one showed any potency, albeit only at high concentrations, even though there is high protein sequence similarity in the CPSF3 protein target compared to other protozoan pathogen species. This finding suggests that benzoxaboroles currently of interest for the treatment of African animal trypanosomiasis, toxoplasmosis, cryptosporidiosis and malaria may not be suitable for the treatment of ECF. We conclude that testing of further benzoxaborole compounds is needed to fully determine whether any lead compounds can be identified to target T. parva.

Sequence similarity network and protein structure prediction offer insights into the evolution of microbial pathways for ferrous iron oxidation.

Dissimilatory ferrous iron [Fe(II)] oxidation is a well-established microbial energy generation strategy. This study aims to comprehensively investigate the distribution and evolution of recognized Fe(II) oxidation pathways through comparative analysis. Interestingly, we have discovered a wide range of taxonomic groups that harbor homologs to known Fe(II) oxidation proteins. The presence of these homologs among phylogenetically distant lineages and their frequent association with mobile genetic elements strongly suggest horizontal gene transfer events involving Fe(II) oxidation proteins, such as the rus operon of Acidithiobacillus and Cyc572 from Leptospirillum lineages belonging to classes Gammaproteobacteria and Betaproteobacteria often present at the hub positions of the protein sequence similarity networks from which homologs of other taxa are derived. In addition, RoseTTAFold predictions have provided valuable insights into the structural characteristics of previously unknown Fe(II) oxidation components. Despite having limited sequence identity, a significant number of acknowledged proteins involved in different Fe(II) oxidation pathways exhibit close structural similarities, including Cyc2 and Cyc572. Collectively, this study significantly enhances our understanding of the distribution and evolution of microbial ferrous iron oxidation pathways. IMPORTANCE Microbial Fe(II) oxidation is a crucial process that harnesses and converts the energy available in Fe, contributing significantly to global element cycling. However, there are still many aspects of this process that remain unexplored. In this study, we utilized a combination of comparative genomics, sequence similarity network analysis, and artificial intelligence-driven structure modeling methods to address the lack of structural information on Fe(II) oxidation proteins and offer a comprehensive perspective on the evolution of Fe(II) oxidation pathways. Our findings suggest that several microbial Fe(II) oxidation pathways currently known may have originated within classes Gammaproteobacteria and Betaproteobacteria.

Molecular Modelling and Structure Analysis of Lectins from Vigna linearis

A study was conducted during the year 2022–2023 for in silico analysis of two Vigna linearis lectins in the computer lab, CDFT, Jodhpur, Rajasthan, India. Amino acid sequences of both lectins were obtained from NCBI database and analyzed using different computational tools. Primary structure analysis using ProtParam server revealed acidic, stable and hydrophobic nature of both VLL proteins. Both proteins were predicted heavily glycosylated with a single signal peptide cleavage site (A26-A27). High sequence similarity of both VLL proteins was observed with adzuki, rice bean and moth bean lectins. Both VLL proteins had high proportions of β-sheets in their secondary structure. Good quality 3D structures for both VLL proteins were modelled using Modeller software. A Jelly roll fold, also known as β-sandwich structure was identified in the 3D structure of both VLL proteins. Profunc server annotated both lectins as a carbohydrate binding regulatory protein with significant role in plant defense (GO terms: 0065007, 0006952, 0050896, 0006950, 0009607, 0030246). Galaxy Site web server predicted galactose, A2G (2-acetamido-2-deoxy-α-D-galactopyranose), MFU (Methyl α-L-fucopyranoside) and adenine as common binding ligands for both VLL proteins. Lig Plot analysis revealed hydrogen bonding and hydrophobic interactions were the major bonding interactions between VLL proteins and putative ligands. The results of the study would provide a base for conducting future research on Vigna linearis lectins for different applications.

HNetGO: protein function prediction via heterogeneous network transformer.

Protein function annotation is one of the most important research topics for revealing the essence of life at molecular level in the post-genome era. Current research shows that integrating multisource data can effectively improve the performance of protein function prediction models. However, the heavy reliance on complex feature engineering and model integration methods limits the development of existing methods. Besides, models based on deep learning only use labeled data in a certain dataset to extract sequence features, thus ignoring a large amount of existing unlabeled sequence data. Here, we propose an end-to-end protein function annotation model named HNetGO, which innovatively uses heterogeneous network to integrate protein sequence similarity and protein-protein interaction network information and combines the pretraining model to extract the semantic features of the protein sequence. In addition, we design an attention-based graph neural network model, which can effectively extract node-level features from heterogeneous networks and predict protein function by measuring the similarity between protein nodes and gene ontology term nodes. Comparative experiments on the human dataset show that HNetGO achieves state-of-the-art performance on cellular component and molecular function branches.

Characterization and Functional Analysis of Chalcone Synthase Genes in Highbush Blueberry (Vaccinium corymbosum).

Chalcone synthase (CHS) is the first key enzyme-catalyzing plant flavonoid biosynthesis. Until now, however, the blueberry CHS gene family has not been systematically characterized and studied. In this study, we identified 22 CHS genes that could be further classified into four subfamilies from the highbush blueberry (Vaccinium corymbosum) genome. This classification was well supported by the high nucleotide and protein sequence similarities and similar gene structure and conserved motifs among VcCHS members from the same subfamily. Gene duplication analysis revealed that the expansion of the blueberry CHS gene family was mainly caused by segmental duplications. Promoter analysis revealed that the promoter regions of VcCHSs contained numerous cis-acting elements responsive to light, phytohormone and stress, along with binding sites for 36 different types of transcription factors. Gene expression analysis revealed that Subfamily I VcCHSs highly expressed in fruits at late ripening stages. Through transient overexpression, we found that three VcCHSs (VcCHS13 from subfamily II; VcCHS8 and VcCHS21 from subfamily I) could significantly enhance the anthocyanin accumulation and up-regulate the expression of flavonoid biosynthetic structural genes in blueberry leaves and apple fruits. Notably, the promoting effect of the Subfamily I member VcCHS21 was the best. The promoter of VcCHS21 contains a G-box (CACGTG) and an E-box sequence, as well as a bHLH binding site. A yeast one hybridization (Y1H) assay revealed that three anthocyanin biosynthesis regulatory bHLHs (VcAN1, VcbHLH1-1 and VcbHLH1-2) could specifically bind to the G-box sequence (CACGTG) in the VcCHS21 promoter, indicating that the expression of VcCHS21 was regulated by bHLHs. Our study will be helpful for understanding the characteristics and functions of blueberry CHSs.

Comparative Structure Analysis of the Multi-Domain, Cell Envelope Proteases of Lactic Acid Bacteria.

Lactic acid bacteria (LAB) have an extracellular proteolytic system that includes a multi-domain, cell envelope protease (CEP) with a subtilisin homologous protease domain. These CEPs have different proteolytic activities despite having similar protein sequences. Structural characterization has previously been limited to CEP homologs of dairy- and human-derived LAB strains, excluding CEPs of plant-derived LAB strains. CEP structures are a challenge to determine experimentally due to their large size and attachment to the cell envelope. This study aims to clarify the prevalence and structural diversity of CEPs by using the structure prediction software AlphaFold 2. Domain boundaries are clarified based on a comparative analysis of 21 three-dimensional structures, revealing novel domain architectures of CEP homologs that are not necessarily restricted to specific LAB species or ecological niches. The C-terminal flanking region of the protease domain is divided into fibronectin type-III-like domains with various structural traits. The analysis also emphasizes the existence of two distinct domains for cell envelope attachment that are preceded by an intrinsically disordered cell wall spanning domain. The domain variants and their combinations provide CEPs with different stability, proteolytic activity, and potentially adhesive properties, making CEPs targets for steering proteolytic activity with relevance for both food development and human health.

A broad inhibitor of acyl-acyl carrier protein synthetases

The acyl-acyl carrier protein synthetase enzyme enables some bacteria to scavenge free fatty acids from the environment for direct use in lipids. This fatty acid recycling pathway can help pathogens circumvent fatty acid synthase (FAS) inhibition with established antibiotics and those in clinical development. AasS enzymes are surprisingly hard to identify as they show high sequence similarity to other adenylate forming enzymes, and only a handful have been correctly annotated to date. Four recently discovered AasS enzymes from Gram negative bacteria, Chlamydia trachomatis, Neisseria gonorrhoeae, and Alistipes finegoldii, form distinct clusters in protein sequence similarity networks and have varying substrate preferences. We previously synthesized C10-AMS, an inhibitor of AasS that mimics the acyl-AMP reaction intermediate. Here we tested its ability to be broadly applicable to enzymes in this class, and found it inhibits all four newly annotated AasS enzymes. C10-AMS therefore provides a tool to study the role of AasS in fatty acid recycling in pathogenic bacteria as well as offers a platform for antibiotic development.

Predicting Drug-Target Interaction Via Self-Supervised Learning.

Recent advances in graph representation learning provide new opportunities for computational drug-target interaction (DTI) prediction. However, it still suffers from deficiencies of dependence on manual labels and vulnerability to attacks. Inspired by the success of self-supervised learning (SSL) algorithms, which can leverage input data itself as supervision,we propose SupDTI, a SSL-enhanced drug-target interaction prediction framework based on a heterogeneous network (i.e., drug-protein, drug-drug, and protein-protein interaction network; drug-disease, drug-side-effect, and protein-disease association network; drug-structure and protein-sequence similarity network). Specifically, SupDTI is an end-to-end learning framework consisting of five components. First, localized and globalized graph convolutions are designed to capture the nodes' information from both local and global perspectives, respectively. Then, we develop a variational autoencoder to constrain the nodes' representation to have desired statistical characteristics. Finally, a unified self-supervised learning strategy is leveraged to enhance the nodes' representation, namely, a contrastive learning module is employed to enable the nodes' representation to fit the graph-level representation, followed by a generative learning module which further maximizes the node-level agreement across the global and local views by learning the probabilistic connectivity distribution of the original heterogeneous network. Experimental results show that our model can achieve better prediction performance than state-of-the-art methods.

Genome-Wide Identification and Expression Analysis of Respiratory Burst Oxidase Homolog (RBOH) Gene Family in Eggplant (Solanum melongena L.) under Abiotic and Biotic Stress.

Respiratory burst oxidase homologs (RBOHs) are important proteins that catalyze the production of reactive oxygen species (ROS), which play important roles in growth and stress response. For a comprehensive analysis of SmRBOH genes, we conducted genome-wide identification of the SmRBOH gene family in eggplant (Solanum melongena L.) and analyzed the expression of SmRBOHs under abiotic (salt, high-temperature, and low-temperature) and biotic stress (Verticillium dahliae inoculation) by quantitative real-time PCR (qRT-PCR). The result showed that a total of eight SmRBOH members were identified from the genome database of eggplant, and they were relatively evenly distributed across seven chromosomes. The analysis of Motif and the conserved domain showed that SmRBOHs have high similarity in protein sequences and functions. Based on phylogenetics, SmRBOHs were classified into three distinct clades. Furthermore, the promoter regions of SmRBOHs were found to contain different cis-elements. Additionally, the results of the qRT-PCR demonstrated differential expression patterns of SmRBOHs in different tissues (the roots, stems, and leaves) and stress conditions. SmRBOHB, SmRBOHD, SmRBOHH1, and SmRBOHH2 showed significant upregulation (>20-fold) under at least one stress condition. Subcellular localization analysis of the above four members further confirmed that they localized on the plasma membrane. This study provides a theoretical foundation for understanding the functions of SmRBOHs in eggplant.

Identification of the zebrafish homologues of IMPG2, a retinal proteoglycan.

Photoreceptor outer segments are surrounded by a carbohydrate-rich matrix, the interphotoreceptor matrix, necessary for physiological retinal function. Few roles for molecules characterizing the interphotoreceptor matrix have been clearly defined. Recent studies have found the presence of nonsense mutations in the interphotoreceptor matrix proteoglycan 2 (IMPG2) gene in patients affected by retinal dystrophies. IMPG2 encodes for a proteoglycan synthesized by photoreceptors and secreted in the interphotoreceptor matrix. Little is known about the structure and function of this protein, we thus decided to characterize zebrafish impg2. In zebrafish there are two Impg2 proteins, Impg2a and Impg2b. We generated a phylogenetic tree based on IMPG2 protein sequence similarity among vertebrates, showing a significant similarity between humans and teleosts. The human and zebrafish proteins share conserved domains, as also shown by homology models. Expression analyses of impg2a and impg2b show a continued expression in the photoreceptor layer starting from developmental stages and continuing through adulthood. Between 1 and 6months post-fertilization, there is a significant shift of Impg2 expression toward the outer segment region, suggesting an increase in secretion. This raises intriguing hypotheses about its possible role(s) during retinal maturation, laying the groundwork for the generation of most needed models for the study of IMPG2-related inherited retinal dystrophies.

Genome-scale metabolic modeling and in silico analysis of opportunistic skin pathogen Cutibacterium acnes.

Cutibacterium acnes, one of the most abundant skin microbes found in the sebaceous gland, is known to contribute to the development of acne vulgaris when its strains become imbalanced. The current limitations of acne treatment using antibiotics have caused an urgent need to develop a systematic strategy for selectively targeting C. acnes, which can be achieved by characterizing their cellular behaviors under various skin environments. To this end, we developed a genome-scale metabolic model (GEM) of virulent C. acnes, iCA843, based on the genome information of a relevant strain from ribotype 5 to comprehensively understand the pathogenic traits of C. acnes in the skin environment. We validated the model qualitatively by demonstrating its accuracy prediction of propionate and acetate production patterns, which were consistent with experimental observations. Additionally, we identified unique biosynthetic pathways for short-chain fatty acids in C. acnes compared to other GEMs of acne-inducing skin pathogens. By conducting constraint-based flux analysis under endogenous carbon sources in human skin, we discovered that the Wood-Werkman cycle is highly activated under acnes-associated skin condition for the regeneration of NAD, resulting in enhanced propionate production. Finally, we proposed potential anti-C. acnes targets by using the model-guided systematic framework based on gene essentiality analysis and protein sequence similarity search with abundant skin microbiome taxa.

Motif2Mol: Prediction of New Active Compounds Based on Sequence Motifs of Ligand Binding Sites in Proteins Using a Biochemical Language Model.

In drug design, the prediction of new active compounds from protein sequence data has only been attempted in a few studies thus far. This prediction task is principally challenging because global protein sequence similarity has strong evolutional and structural implications, but is often only vaguely related to ligand binding. Deep language models adapted from natural language processing offer new opportunities to attempt such predictions via machine translation by directly relating amino acid sequences and chemical structures to each based on textual molecular representations. Herein, we introduce a biochemical language model with transformer architecture for the prediction of new active compounds from sequence motifs of ligand binding sites. In a proof-of-concept application on inhibitors of more than 200 human kinases, the Motif2Mol model revealed promising learning characteristics and an unprecedented ability to consistently reproduce known inhibitors of different kinases.