SoluProt: prediction of soluble protein expression in<i>Escherichia coli</i>

Hon, Jiří; Marusiak, Martin; Martínek, Tomáš; Kunka, Antonín; Zendulka, Jaroslav; Bednář, David; Damborský, Jiřı́

doi:10.1093/bioinformatics/btaa1102

Cited by 111 publications

(113 citation statements)

References 47 publications

Supporting

Mentioning

113

Contrasting

Order By: Relevance

“…Soluble expression of recombinant proteins in E. coli was predicted by Soluprot 1.0 (40). Transmembrane helices were predicted with TMHMM 2.0 (41).…”

Section: Detection Of Exposed Proteinsmentioning

confidence: 99%

Designing Multi-Antigen Vaccines Against Acinetobacter baumannii Using Systemic Approaches

McConnell

Martín-Galiano

2021

Front. Immunol.

View full text Add to dashboard Cite

Vaccines and monoclonal antibodies are promising approaches for preventing and treating infections caused by multidrug resistant Acinetobacter baumannii. However, only partial protection has been achieved with many previously tested protein antigens, which suggests that vaccines incorporating multiple antigens may be necessary in order to obtain high levels of protection. Several aspects that use the wealth of omic data available for A. baumannii have not been fully exploited for antigen identification. In this study, the use of fractionated proteomic and computational data from ~4,200 genomes increased the number of proteins potentially accessible to the humoral response to 8,824 non-redundant proteins in the A. baumannii panproteome. Among them, 59% carried predicted B-cell epitopes and T-cell epitopes recognized by two or more alleles of the HLA class II DP supertype. Potential cross-reactivity with human proteins was detected for 8.9% of antigens at the protein level and 2.7% at the B-cell epitope level. Individual antigens were associated with different infection types by genomic, transcriptomic or functional analyses. High intra-clonal genome density permitted the identification of international clone II as a “vaccitype”, in which 20% of identified antigens were specific to this clone. Network-based centrality measurements were used to identify multiple immunologic nodes. Data were formatted, unified and stored in a data warehouse database, which was subsequently used to identify synergistic antigen combinations for different vaccination strategies. This study supports the idea that integration of multi-omic data and fundamental knowledge of the pathobiology of drug-resistant bacteria can facilitate the development of effective multi-antigen vaccines against these challenging infections.

show abstract

“…Soluble expression of recombinant proteins in E. coli was predicted by Soluprot 1.0 (40). Transmembrane helices were predicted with TMHMM 2.0 (41).…”

Section: Detection Of Exposed Proteinsmentioning

confidence: 99%

Designing Multi-Antigen Vaccines Against Acinetobacter baumannii Using Systemic Approaches

McConnell

Martín-Galiano

2021

Front. Immunol.

View full text Add to dashboard Cite

show abstract

“…The North East Structural Consortium (NESG) expressed 9644 proteins in E. coli using a unified production pipeline (Price et al, 2011) and provide integer scores (0-5) for both expression (E) and solubility (S). The proteins are part of the TargetTrack database, but the scores were obtained by Hon et al [2021] from the original authors. We remove sequences that have multiple scores and use the remaining 9272 sequences in two ways.…”

Section: Price Datasetmentioning

confidence: 99%

“…The PaRSnIP (Rawi et al, 2017), DeepSol (Khurana et al, 2018), and SKADE (Raimondi et al, 2020) soluble expression predictors were built using the curated train set and were shown to achieve high scores on the test set. However, it was noticed that these tools generalize poorly (Bhandari et al, 2020, Hon et al, 2021). Raimondi et al [2020] showed that the SKADE model focused mostly on the N- and C- termini and validated that DeepSol did the same using an experiment that involved cropping the starting and ending segments of the sequences.…”

Section: Datamentioning

confidence: 99%

NetSolP: predicting protein solubility in E. coli using language models

Thumuluri

Martiny

Armenteros

et al. 2021

Preprint

View full text Add to dashboard Cite

Solubility and expression levels of proteins can be a limiting factor for large-scale studies and industrial production. By determining the solubility and expression directly from the protein sequence, the success rate of wet-lab experiments can be increased. In this study, we focus on predicting the solubility and usability for purification of proteins expressed in Escherichia coli directly from the sequence. Our model NetSolP is based on deep-learning protein language models called transformers and we show that it achieves state-of-the-art performance and improves extrapolation across datasets. As we find current methods are built on biased datasets, we curate existing datasets by using strict sequence-identity partitioning and ensure that there is minimal bias in the sequences.

show abstract

“…Instead of using the trial-and-error approach to get enough protein overexpression, tools that can direct the selection of genes with a higher probability of successful overexpression are desirable. Several tools have been developed for the prediction of soluble overexpression in Escherichia coli, including PROSO II (Smialowski et al, 2012), PaRSnIP (Rawi et al, 2018), DeepSol (Khurana et al, 2018), SKADE (Raimondi et al, 2020), and SoluProt (Hon et al, 2021). In addition, some tools exist for the more specific prediction of solubility, which is an important element in soluble preotein expression.…”

Section: Introductionmentioning

confidence: 99%

“…These include Protein-Sol (Hebditch et al, 2017) and SoDoPe (Bhandari et al, 2020). The mentioned tools use the primary structure as input and calculate various sequence-based features (e.g., hydrophobicity, charge, kmer frequencies, disorder), and they use various machine learning techniques: support vector machines (Agostini et al, 2014), gradient boosting machines (Rawi et al, 2018;Hon et al, 2021), neural networks (Khurana et al, 2018;Raimondi et al, 2020), or other statistical methods (Smialowski et al, 2012;Hebditch et al, 2017;Bhandari et al, 2020). However, all these tools (with the exception of Protein-Sol) have been developed especially with the host Escherichia coli in mind, and it is an open question whether their results can be generalized to other production organisms.…”

Section: Introductionmentioning

confidence: 99%

Deep protein representations enable recombinant protein expression prediction

Martiny¹,

Armenteros

Salomon

et al. 2021

Preprint

View full text Add to dashboard Cite

A crucial process in the production of industrial enzymes is recombinant gene expression, which aims to induce enzyme overexpression of the genes in a host microbe. Current approaches for securing overexpression rely on molecular tools such as adjusting the recombinant expression vector, adjusting cultivation conditions, or performing codon optimizations. However, such strategies are time-consuming, and an alternative strategy would be to select genes for better compatibility with the recombinant host. Several methods for predicting expressibility and solubility are available; however, they are all optimized for the expression host Escherichia coli. We show that these tools are not suited for predicting expression potential in the industrially important host Bacillus subtilis. Instead, we build a B. subtilis-specific machine learning model for expressibility prediction. Given millions of unlabelled proteins, and a small labelled dataset, we can successfully train such a predictive model. The unlabelled proteins provide a performance boost relative to using amino acid frequencies of the labelled proteins as input. On average, we obtain a modest performance of 0.64 area-under-the-curve (AUC) and 0.2 Matthews correlation coeffcient (MCC). However, we find that this is sufficient to be useful for prioritization of expression candidates. Moreover, the predicted class probabilities are correlated with expression levels. A number of features related to protein expression, including base frequencies and solubility, are captured by the model.

show abstract

SoluProt: prediction of soluble protein expression inEscherichia coli

Cited by 111 publications

References 47 publications

Designing Multi-Antigen Vaccines Against Acinetobacter baumannii Using Systemic Approaches

Designing Multi-Antigen Vaccines Against Acinetobacter baumannii Using Systemic Approaches

NetSolP: predicting protein solubility in E. coli using language models

Deep protein representations enable recombinant protein expression prediction

Contact Info

Product

Resources

About