BUSCA: an integrative web server to predict subcellular localization of proteins

Savojardo, Castrense; Martelli, Pier Luigi; Fariselli, Piero; Profiti, Giuseppe; Casadio, Rita

doi:10.1093/nar/gky320

Cited by 330 publications

(236 citation statements)

References 26 publications

(37 reference statements)

Supporting

Mentioning

232

Contrasting

Order By: Relevance

“…The resulting collection of protein sequences was used to generate a consensus sequence for each locus for N. meningitidis and N. gonorrhoeae separately. The consensus protein sequences for NmB and N. gonorrhoeae were provided as input to 3 subcellular localization Web servers, Cello2Go (67), BUSCA (68), and PSORTb (69). The molecular function of the proteins was predicted using InterProScan (70) software.…”

Section: Methodsmentioning

confidence: 99%

Genetic Similarity of Gonococcal Homologs to Meningococcal Outer Membrane Proteins of Serogroup B Vaccine

Marjuki

Topaz

Joseph

et al. 2019

mBio

View full text Add to dashboard Cite

The human pathogens Neisseria gonorrhoeae and Neisseria meningitidis share high genome identity. Retrospective analysis of surveillance data from New Zealand indicates the potential cross-protective effect of outer membrane vesicle (OMV) meningococcal serogroup B vaccine (MeNZB) against N. gonorrhoeae. A licensed OMV-based MenB vaccine, MenB-4C, consists of a recombinant FHbp, NhbA, NadA, and the MeNZB OMV. Previous work has identified several abundantly expressed outer membrane proteins (OMPs) as major components of the MenB-4C OMV with high sequence similarity between N. gonorrhoeae and N. meningitidis, suggesting a mechanism for cross-protection. To build off these findings, we performed comparative genomic analysis on 970 recent N. gonorrhoeae isolates collected through a U.S surveillance system against N. meningitidis serogroup B (NmB) reference sequences. We identified 1,525 proteins that were common to both Neisseria species, of which 57 proteins were predicted to be OMPs using in silico methods. Among the MenB-4C antigens, NhbA showed moderate sequence identity (73%) to the respective gonococcal homolog, was highly conserved within N. gonorrhoeae, and was predicted to be surface expressed. In contrast, the gonococcal FHbp was predicted not to be surface expressed, while NadA was absent in all N. gonorrhoeae isolates. Our work confirmed recent observations (E. A. Semchenko, A. Tan, R. Borrow, and K. L. Seib, Clin Infect Dis, 2018, https://doi.org/10.1093/cid/ciy1061) and describes homologous OMPs from a large panel of epidemiologically relevant N. gonorrhoeae strains in the United States against NmB reference strains. Based on our results, we report a set of OMPs that may contribute to the previously observed cross-protection and provide potential antigen targets to guide the next steps in gonorrhea vaccine development. IMPORTANCE Gonorrhea, a sexually transmitted disease, causes substantial global morbidity and economic burden. New prevention and control measures for this disease are urgently needed, as strains resistant to almost all classes of antibiotics available for treatment have emerged. Previous reports demonstrate that cross-protection from gonococcal infections may be conferred by meningococcal serogroup B (MenB) outer membrane vesicle (OMV)-based vaccines. Among 1,525 common proteins shared across the genomes of both N. gonorrhoeae and N. meningitidis, 57 proteins were predicted to be surface expressed (outer membrane proteins [OMPs]) and thus preferred targets for vaccine development. The majority of these OMPs showed high sequence identity between the 2 bacterial species. Our results provide valuable insight into the meningococcal antigens present in the current OMV-containing MenB-4C vaccine that may contribute to cross-protection against gonorrhea and may inform next steps in gonorrhea vaccine development.

show abstract

Section: Methodsmentioning

confidence: 99%

Genetic Similarity of Gonococcal Homologs to Meningococcal Outer Membrane Proteins of Serogroup B Vaccine

Marjuki

Topaz

Joseph

et al. 2019

mBio

View full text Add to dashboard Cite

show abstract

“…2a). However, a recent publication might suggest this view to be possibly slightly distorted by the choice of data set and number of states (Savojardo, et al, 2018). Another reason for this result could be attributed to the technical limitation forcing proteins to be chopped to 1024 residues (using only the first and last 512 residues for protein longer than 512 amino acids).…”

Section: Discussionmentioning

confidence: 99%

End-to-end multitask learning, from protein language to protein features without alignments

Elnaggar

Heinzinger

Dallago

et al. 2019

Preprint

View full text Add to dashboard Cite

17Correctly predicting features of protein structure and function from amino acid sequence alone 18 remains a supreme challenge for computational biology. For almost three decades, state-of-19 the-art approaches combined machine learning and evolutionary information from multiple 20 sequence alignments. Exponentially growing sequence databases make it infeasible to gather 21 evolutionary information for entire microbiomes or meta-proteomics. On top, for many 22 important proteins (e.g. dark proteome and intrinsically disordered proteins) evolutionary 23 information remains limited. Here, we introduced a novel approach combining recent advances 24 of Language Models (LMs) with multi-task learning to successfully predict aspects of protein 25 structure (secondary structure) and function (cellular component or subcellular localization) 26 without using any evolutionary information from alignments. Our approach fused self-27 supervised pre-training LMs on an unlabeled big dataset (UniRef50, corresponding to 9.6 28 billion words) with supervised training on labelled high-quality data in one single end-to-end 29 network. We provided a proof-of-principle for the novel concept through the semi-successful 30 per-residue prediction of protein secondary structure and through per-protein predictions of 31 localization (Q10=69%) and the distinction between integral membrane and water-soluble 32 proteins (Q2=89%). Although these results did not reach the levels obtained by the best 33 available methods using evolutionary information from alignments, these less accurate multi-34 task predictions have the advantage of speed: they are 300-3000 times faster (where HHblits 35 needs 30-300 seconds on average, our method needed 0.045 seconds). These new results 36 push the boundaries of predictability towards grayer and darker areas of the protein space, 37 allowing to make reliable predictions for proteins which were not accessible by previous 38 methods. On top, our method remains scalable as it removes the necessity to search sequence 39 databases for evolutionary related proteins. 40 41

show abstract

“…21 ). Sub-cellular localisation of CAZymes was predicted using BUSCA web 63 . To link the degradation of the different lignocellulose fractions and subsequent sugar utilisation, we looked for the presence of suitable sugar transporters and also specific sugar isomerases and kinases.…”

Section: Methodsmentioning

confidence: 99%

Targeted biomass degradation by the higher termite gut system - integrative omics applied to host and its gut microbiome

Całusińska

Marynowska

Bertucci

et al. 2020

Preprint

View full text Add to dashboard Cite

26Miscanthus sp. is regarded as suitable biomass for different biorefinery value chains. However, due 27 to high recalcitrance, its wide use is yet untapped. Termite is the most efficient lignocellulose 28 degrading insect, and its success results from synergistic cooperation with its gut microbiome. 29Here, we investigated at holobiont level the dynamic adaptation of a higher termite Cortaritermes 30 sp. to imposed Miscanthus diet, with a long-term objective of overcoming lignocellulose 31 recalcitrance. We used an integrative omics approach, comprising amplicon sequencing, 32 metagenomics and metatranscriptomics that we combined with enzymatic characterisation of 33 carbohydrate active enzymes from termite gut Fibrobacteres and Spirochaetae. Adaptation to the 34 new diet was evidenced by reduced gut bacterial diversity and modified gene expression profiles, 35 further suggesting a shift towards utilisation of cellulose and arabinoxylan, two main components of 36Miscanthus lignocellulose. Low identity of reconstructed microbial genomes to microbes from 37 closely related termite species, supported the hypothesis of a strong phylogenetic relationship 38 between host and its gut microbiome. Application-wise, this makes each termite gut system an 39 endless source of enzymes that are potentially industrially relevant. 40This study provides a framework for better understanding the complex lignocellulose degradation by 41 the higher termite gut system and paves a road towards its future bioprospecting. 42 43

show abstract

BUSCA: an integrative web server to predict subcellular localization of proteins

Cited by 330 publications

References 26 publications

Genetic Similarity of Gonococcal Homologs to Meningococcal Outer Membrane Proteins of Serogroup B Vaccine

Genetic Similarity of Gonococcal Homologs to Meningococcal Outer Membrane Proteins of Serogroup B Vaccine

End-to-end multitask learning, from protein language to protein features without alignments

Targeted biomass degradation by the higher termite gut system - integrative omics applied to host and its gut microbiome

Contact Info

Product

Resources

About