Computational prediction of secreted proteins in gram-negative bacteria

Hui, Xinjie; Chen, Zewei; Zhang, Junya; Lu, Moyang; Cai, Xuxia; Deng, Yuping; Hu, Yueming; Wang, Yejun

doi:10.1016/j.csbj.2021.03.019

Cited by 31 publications

(37 citation statements)

References 305 publications

(252 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Currently, there is still a lack of computational methods predicting T1SEs ( Hui et al, 2021 ). Although Luo et al (2015) developed a random forest predictor, the tool or codes were not publically available and therefore a direct comparison could not be performed.…”

Section: Discussionmentioning

confidence: 99%

“…Since the first T1SS substrate, hemolysin A (HlyA), was discovered in 1979 and its nucleotide sequence was determined in 1985 ( Noegel et al, 1979 ; Felmlee et al, 1985 ), the structural characteristics and function of T1SEs have been studied extensively. Typical T1SEs can be classified into three classes simply according to their T1SS ABC transporter types: C39-containing ABC transporters with hydrolase activity, C39-like domain (CLD)–containing ABC transporters without hydrolase activity, and a third type of ABC transporters without any additional N-terminal domain ( Hui et al, 2021 ). Class 1 T1SEs, known as the smallest T1SS substrates, normally contain N-terminal leader peptides.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

T1SEstacker: A Tri-Layer Stacking Model Effectively Predicts Bacterial Type 1 Secreted Proteins Based on C-Terminal Non-repeats-in-Toxin-Motif Sequence Features

et al. 2022

Self Cite

View full text Add to dashboard Cite

Type 1 secretion systems play important roles in pathogenicity of Gram-negative bacteria. However, the substrate secretion mechanism remains largely unknown. In this research, we observed the sequence features of repeats-in-toxin (RTX) proteins, a major class of type 1 secreted effectors (T1SEs). We found striking non-RTX-motif amino acid composition patterns at the C termini, most typically exemplified by the enriched “[FLI][VAI]” at the most C-terminal two positions. Machine-learning models, including deep-learning ones, were trained using these sequence-based non-RTX-motif features and further combined into a tri-layer stacking model, T1SEstacker, which predicted the RTX proteins accurately, with a fivefold cross-validated sensitivity of ∼0.89 at the specificity of ∼0.94. Besides substrates with RTX motifs, T1SEstacker can also well distinguish non-RTX-motif T1SEs, further suggesting their potential existence of common secretion signals. T1SEstacker was applied to predict T1SEs from the genomes of representative Salmonella strains, and we found that both the number and composition of T1SEs varied among strains. The number of T1SEs is estimated to reach 100 or more in each strain, much larger than what we expected. In summary, we made comprehensive sequence analysis on the type 1 secreted RTX proteins, identified common sequence-based features at the C termini, and developed a stacking model that can predict type 1 secreted proteins accurately.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

T1SEstacker: A Tri-Layer Stacking Model Effectively Predicts Bacterial Type 1 Secreted Proteins Based on C-Terminal Non-repeats-in-Toxin-Motif Sequence Features

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…Currently, there is still a lack of computational methods predicting T1SEs (Hui et al 2021). Although Luo et al developed a random forest predictor, the tool or codes were not publically available and therefore a direct comparison could not be performed (Luo et al, 2015).…”

Section: Discussionmentioning

confidence: 99%

“…T1SEs can be divided into 3 classes (Classes 1~3) according to different ABC transporters types: C39-containing ABC transporters with hydrolase activity, C39-like domain (CLD)-containing ABC transporters without hydrolase activity, and a third type of ABC transporters without any additional N-terminal domain (Hui et al, 2021). Class 1 T1SEs, known as the smallest T1SS substrates, normally contain N-terminal leader peptides.…”

Section: Introductionmentioning

confidence: 99%

T1SEstacker: A tri-layer stacking model effectively predicts bacterial type 1 secreted proteins based on C-terminal non-RTX-motif sequence features

Chen

Zhao

Hui

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

The proteins secreted through type 1 secretion systems often play important roles in pathogenicity of various gram-negative bacteria. However, the type 1 secretion mechanism remains unknown. In this research, we observed the sequence features of RTX proteins, a major class of type 1 secreted substrates. We found striking non-RTX-motif amino acid composition patterns at the C-termini, most typically exemplified by the enriched ‘[FLI][VAI]’ at the most C-terminal two positions. Machine-learning models, including deep-learning models, were trained using these sequence-based non-RTX-motif features, and further combined into a tri-layer stacking model, T1SEstacker, which predicted the RTX proteins accurately, with a 5-fold cross-validated sensitivity of ~0.89 at the specificity of ~0.94. Besides substrates with RTX motifs, T1SEstacker can also well distinguish non-RTX-motif type 1 secreted proteins, further suggesting their potential existence of common secretion signals. In summary, we made comprehensive sequence analysis on the type 1 secreted RTX proteins, identified common sequence-based features at the C-termini, and developed a stacking model that can predict type 1 secreted proteins accurately.

show abstract

“…We also considered whether the second and third amino acids might facilitate protein secretion. There are several pathways for protein translocation into or across the plasma membrane that can depend on the physical properties of the amino acids of 5-30 amino acid secretory signal peptides (SPs), usually located near the amino terminus 26,27 (recently reviewed by Owji et al, 2018 and Hui et al, 2021) 28,29 . But the details are too complex to be unraveled by our analysis of average features.…”

Section: Figure 5 Amino Acid Bias Near Amino-termini Amino Acids Are Rank-ordered By the Frequencies At Internal Positions (Blue Bars) Tomentioning

confidence: 99%

Analyses of the 5’ Ends of Escherichia coli ORFs

Ward

Curran

2021

Preprint

View full text Add to dashboard Cite

Sequence biases at 5-prime ends of coding sequences differ from those of the remainder of ORFs, reflecting differences in function. Internal sequence biases promote translational efficiency by several mechanisms including correlating codon usage and tRNA concentration. However, the early region may also facilitate translational initiation, establishment of the reading frame, and polypeptide processing. Here we examine the beginnings of the ORFs of an Escherichia coli K12 reference genome. The results extend previous observations of A-richness to include an overabundance of the AAA triplet in all reading frames, consistent with the hypothesis that the beginnings of ORFs contribute to initiation site accessibility. Results are also consistent with the idea that the first two amino acids are under selection because they facilitate solvation of the amino-terminus at the end of the ribosomal exit channel. Moreover, serine is highly overrepresented as the second amino acid, possibly because it can facilitate removal of the terminal formylmethionine. Non-AUG initiation codons are known to be less efficient than AUG at directing initiation, presumably because of relatively weak base pairing to the initiator-tRNA. But non-UAG initiation codons are not followed by unusual 3-prime nearest neighbor codons. Moreover, the four NUG initiation codons do not differ in their propensity to frameshift in an assay known to be sensitive to base pair strength. Altogether, these data suggest that the 5-prime ends of ORFs are under selection for several functions, and that initiation codon identity may not critical beyond its role in initiation.

show abstract

Computational prediction of secreted proteins in gram-negative bacteria

Abstract: Graphical abstract

Cited by 31 publications

References 305 publications

T1SEstacker: A Tri-Layer Stacking Model Effectively Predicts Bacterial Type 1 Secreted Proteins Based on C-Terminal Non-repeats-in-Toxin-Motif Sequence Features

T1SEstacker: A Tri-Layer Stacking Model Effectively Predicts Bacterial Type 1 Secreted Proteins Based on C-Terminal Non-repeats-in-Toxin-Motif Sequence Features

T1SEstacker: A tri-layer stacking model effectively predicts bacterial type 1 secreted proteins based on C-terminal non-RTX-motif sequence features

Analyses of the 5’ Ends of Escherichia coli ORFs

Contact Info

Product

Resources

About