Limits...
RNA-Seq for Plant Pathogenic Bacteria.

Kimbrel JA, Di Y, Cumbie JS, Chang JH - Genes (Basel) (2011)

Bottom Line: The throughput and single-base resolution of RNA-Sequencing (RNA-Seq) have contributed to a dramatic change in transcriptomic-based inquiries and resulted in many new insights into the complexities of bacterial transcriptomes.RNA-Seq could contribute to similar advances in our understanding of plant pathogenic bacteria but it is still a technology under development with limitations and unknowns that need to be considered.We also discuss the technical and statistical challenges in the practical application of RNA-Seq for studying bacterial transcriptomes and describe some of the currently available solutions.

View Article: PubMed Central - PubMed

Affiliation: Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331, USA. kimbreje@science.oregonstate.edu.

ABSTRACT
The throughput and single-base resolution of RNA-Sequencing (RNA-Seq) have contributed to a dramatic change in transcriptomic-based inquiries and resulted in many new insights into the complexities of bacterial transcriptomes. RNA-Seq could contribute to similar advances in our understanding of plant pathogenic bacteria but it is still a technology under development with limitations and unknowns that need to be considered. Here, we review some new developments for RNA-Seq and highlight recent findings for host-associated bacteria. We also discuss the technical and statistical challenges in the practical application of RNA-Seq for studying bacterial transcriptomes and describe some of the currently available solutions.

No MeSH data available.


Related in: MedlinePlus

Differential expression as a function of transcript length. RNA-Seq data of transcriptomes from Arabidopsis thaliana infected with nonpathogenic bacteria or mock inoculated were analyzed using the GENE-counter pipeline configured with the NBPSeq package. (A) The differentially induced genes (y-axis) were binned based on equal range of transcript lengths (x-axis). A regression line is plotted. (B) Expressed genes from all replicates from both treatments are represented as a percentage within each bin defined based on equal range of transcript length.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3927590&req=5

f3-genes-02-00689: Differential expression as a function of transcript length. RNA-Seq data of transcriptomes from Arabidopsis thaliana infected with nonpathogenic bacteria or mock inoculated were analyzed using the GENE-counter pipeline configured with the NBPSeq package. (A) The differentially induced genes (y-axis) were binned based on equal range of transcript lengths (x-axis). A regression line is plotted. (B) Expressed genes from all replicates from both treatments are represented as a percentage within each bin defined based on equal range of transcript length.

Mentions: In experiments where gene expression is being compared between treatment groups, the variability due to differences in transcript lengths and other technical biases that we have not discussed, are less of an issue, since they presumably affect the same genes to the same degree across different treatment groups. The same cannot be said for other types of analyses that rely on direct or indirect comparisons of expression of a set of genes, such as network or pathway analyses, systems studies, and analysis for enriched gene ontology (GO) terms. Since tests for differential expression are usually more powerful for genes encoding longer transcripts, tests for sets of enriched and differentially expressed genes may be biased towards those that are on average longer in length [65]. To address this issue, a weighted sampling method has been proposed to compensate for length differences [66]. We note, however, that in the original study, the problem of overdispersion was not well understood and some of the data examples that were characterized did not include biological replicates [65]. When we used NBPSeq to identify differentially induced genes from an RNA-Seq dataset comparing transcriptome changes of a host plant challenged with bacteria versus a mock inoculation, we did not observe substantial correlations between differential expression and transcript length (Figure 3) [67]. We feel that further study is needed to fully appreciate the scope and severity of this so-called “length-bias” issue.


RNA-Seq for Plant Pathogenic Bacteria.

Kimbrel JA, Di Y, Cumbie JS, Chang JH - Genes (Basel) (2011)

Differential expression as a function of transcript length. RNA-Seq data of transcriptomes from Arabidopsis thaliana infected with nonpathogenic bacteria or mock inoculated were analyzed using the GENE-counter pipeline configured with the NBPSeq package. (A) The differentially induced genes (y-axis) were binned based on equal range of transcript lengths (x-axis). A regression line is plotted. (B) Expressed genes from all replicates from both treatments are represented as a percentage within each bin defined based on equal range of transcript length.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3927590&req=5

f3-genes-02-00689: Differential expression as a function of transcript length. RNA-Seq data of transcriptomes from Arabidopsis thaliana infected with nonpathogenic bacteria or mock inoculated were analyzed using the GENE-counter pipeline configured with the NBPSeq package. (A) The differentially induced genes (y-axis) were binned based on equal range of transcript lengths (x-axis). A regression line is plotted. (B) Expressed genes from all replicates from both treatments are represented as a percentage within each bin defined based on equal range of transcript length.
Mentions: In experiments where gene expression is being compared between treatment groups, the variability due to differences in transcript lengths and other technical biases that we have not discussed, are less of an issue, since they presumably affect the same genes to the same degree across different treatment groups. The same cannot be said for other types of analyses that rely on direct or indirect comparisons of expression of a set of genes, such as network or pathway analyses, systems studies, and analysis for enriched gene ontology (GO) terms. Since tests for differential expression are usually more powerful for genes encoding longer transcripts, tests for sets of enriched and differentially expressed genes may be biased towards those that are on average longer in length [65]. To address this issue, a weighted sampling method has been proposed to compensate for length differences [66]. We note, however, that in the original study, the problem of overdispersion was not well understood and some of the data examples that were characterized did not include biological replicates [65]. When we used NBPSeq to identify differentially induced genes from an RNA-Seq dataset comparing transcriptome changes of a host plant challenged with bacteria versus a mock inoculation, we did not observe substantial correlations between differential expression and transcript length (Figure 3) [67]. We feel that further study is needed to fully appreciate the scope and severity of this so-called “length-bias” issue.

Bottom Line: The throughput and single-base resolution of RNA-Sequencing (RNA-Seq) have contributed to a dramatic change in transcriptomic-based inquiries and resulted in many new insights into the complexities of bacterial transcriptomes.RNA-Seq could contribute to similar advances in our understanding of plant pathogenic bacteria but it is still a technology under development with limitations and unknowns that need to be considered.We also discuss the technical and statistical challenges in the practical application of RNA-Seq for studying bacterial transcriptomes and describe some of the currently available solutions.

View Article: PubMed Central - PubMed

Affiliation: Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331, USA. kimbreje@science.oregonstate.edu.

ABSTRACT
The throughput and single-base resolution of RNA-Sequencing (RNA-Seq) have contributed to a dramatic change in transcriptomic-based inquiries and resulted in many new insights into the complexities of bacterial transcriptomes. RNA-Seq could contribute to similar advances in our understanding of plant pathogenic bacteria but it is still a technology under development with limitations and unknowns that need to be considered. Here, we review some new developments for RNA-Seq and highlight recent findings for host-associated bacteria. We also discuss the technical and statistical challenges in the practical application of RNA-Seq for studying bacterial transcriptomes and describe some of the currently available solutions.

No MeSH data available.


Related in: MedlinePlus