Genome-wide identification and characterization of small RNAs in <i>Rhodobacter capsulatus</i> and identification of small RNAs affected by loss of the response regulator CtrA

Grüll, Marc P.; Peña‐Castillo, Lourdes; Mulligan, Martin E.; Lang, Andrew S.

doi:10.1080/15476286.2017.1306175

“…Published positive instances of bona fide sRNAs were collected for R. capsulatus (Grüll et al, 2017), included, in addition to four experimentally verified sRNAs, 41 homologous sRNAs (i.e., sRNAs that 69 have high sequence similarity to known sRNAs in other bacterial species or were found to be conserved 70 in the genome of at least two other bacterial species). We randomly selected 80% of the positive instances 71 for training, while setting aside the other 20% for validating the models.…”

mentioning

confidence: 99%

Prioritizing bona fide bacterial small RNAs with machine learning classifiers

Eppenhof

¹

,

Peña‐Castillo

²

2018

Preprint

Self Cite

View full text Add to dashboard Cite

Bacterial small non-coding RNAs (sRNAs) are involved in the control of several cellular processes. Hundreds of putative sRNAs have been identified in many bacterial species through RNA sequencing. The existence of putative sRNAs is usually validated by Northern blot analysis. However, the large amount of novel putative sRNAs reported in the literature makes it impractical to validate in the wet lab each of them. In this work, we applied five machine learning approaches to construct twenty models to discriminate bona fide sRNAs from random genomic sequences in five bacterial species. Sequences were represented using seven features including free energy of their predicted secondary structure, their distances to the closest predicted promoter site and Rho-independent terminator, and their distance to the closest open reading frames (ORFs). To automatically calculate these features, we developed an sRNA Characterization Pipeline (sRNACharP). All sevens features used in the classification task contributed positively to the performance of the predictive models. The five best performing models obtained a median precision of 100% at 10% recall and of 60% at 40% recall across all five bacterial species. Our results suggest that even though there is limited sRNA sequence conservation across different bacterial species, there are intrinsic features of sRNAs that are conserved across taxa. We show that these features are exploited by machine learning approaches to learn a speciesindependent model to prioritize bona fide bacterial sRNAs. Bacterial small non-coding RNAs (sRNAs) are involved in the control of several cellular processes. Hundreds of putative sRNAs have been identified in many bacterial species through RNA sequencing. The existence of putative sRNAs is usually validated by Northern blot analysis. However, the large amount of novel putative sRNAs reported in the literature makes it impractical to validate in the wet lab each of them. In this work, we applied five machine learning approaches to construct twenty models to discriminate bona fide sRNAs from random genomic sequences in five bacterial species. Sequences were represented using seven features including free energy of their predicted secondary structure, their distances to the closest predicted promoter site and Rho-independent terminator, and their distance to the closest open reading frames (ORFs). To automatically calculate these features, we developed an sRNA Characterization Pipeline (sRNACharP). All sevens features used in the classification task contributed positively to the performance of the predictive models. The five best performing models obtained a median precision of 100% at 10% recall and of 60% at 40% recall across all five bacterial species. Our results suggest that even though there is limited sRNA sequence conservation across different bacterial species, there are intrinsic features of sRNAs that are conserved across taxa. We show that these features are exploited by machine learning approaches to learn a species-independent model to pr...

show abstract

“…Published positive instances of bona fide sRNAs were collected for R. capsulatus (Grüll et al, 2017), S. pyogenes (Le Rhun et al, 2016), and S. enterica (Kröger et al, 2012). S. pyogenes and S. enterica positive instances have all been verified by Northern blot analysis; while, R. capsulatus positive instances included, in addition to four experimentally verified sRNAs, 41 homologous sRNAs (i.e., sRNAs that have high sequence similarity to known sRNAs in other bacterial species or were found to be conserved in the genome of at least two other bacterial species).…”

Section: Data Setsmentioning

confidence: 99%

“…Each sRNA is represented as a vector of seven numerical features or attributes, as in Grüll et al (2017).…”

Section: Trainingmentioning

confidence: 99%

“…Bacterial small non-coding RNAs (sRNAs) are regulatory RNAs (usually between 50 to 250 nucleotides) that are known to play a role in the control of several cellular processes (Storz et al, 2011;Michaux et al, 2014). A multitude of putative sRNAs has been identified in many bacterial species through RNA sequencing (e.g., Grüll et al (2017); Thomason et al (2015); Zeng and Sundin (2014); McClure et al…”

Section: Introductionmentioning

confidence: 99%

“…To do that, we need to computationally prioritize sRNAs based on their likelihood of being bona fide sRNAs. As the inter-species sequence conservation of sRNAs is very limited and most sRNAs are species-specific (Gómez-Lozano et al, 2015;Grüll et al, 2017), sRNA prioritization based on sequence similarity to known sRNAs has a low recall rate. However, predictive models generated by machine learning approaches may be able to detect intrinsic features of sRNA sequences common to a number of bacterial species.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Untitled

Supplemental Information 1: Positive and Negative sRNA Instances of Five Bacterial Species

0

View full text Add to dashboard Cite

Bacterial small non-coding RNAs (sRNAs) are involved in the control of several cellular processes. Hundreds of putative sRNAs have been identified in many bacterial species through RNA sequencing. The existence of putative sRNAs is usually validated by Northern blot analysis. However, the large amount of novel putative sRNAs reported in the literature makes it impractical to validate in the wet lab each of them. In this work, we applied five machine learning approaches to construct twenty models to discriminate bona fide sRNAs from random genomic sequences in five bacterial species. Sequences were represented using seven features including free energy of their predicted secondary structure, their distances to the closest predicted promoter site and Rho-independent terminator, and their distance to the closest open reading frames (ORFs). To automatically calculate these features, we developed an sRNA Characterization Pipeline (sRNACharP). All sevens features used in the classification task contributed positively to the performance of the predictive models. The five best performing models obtained a median precision of 100% at 10% recall and of 60% at 40% recall across all five bacterial species. Our results suggest that even though there is limited sRNA sequence conservation across different bacterial species, there are intrinsic features of sRNAs that are conserved across taxa. We show that these features are exploited by machine learning approaches to learn a speciesindependent model to prioritize bona fide bacterial sRNAs.

show abstract