Limits...
Ratio-based analysis of differential mRNA processing and expression of a polyadenylation factor mutant pcfs4 using arabidopsis tiling microarray.

Zheng J, Xing D, Wu X, Shen Y, Kroll DM, Ji G, Li QQ - PLoS ONE (2011)

Bottom Line: Quantitative PCR analysis of a set of DPGs confirmed that most of these genes were truly differentially processed in pcfs4 mutant plants.The enriched GO term "regulation of flower development" among PCFS4 targets further indicated the efficacy of the RADPRE pipeline.This simple but effective program is available upon request.

View Article: PubMed Central - PubMed

Affiliation: Department of Automation, Xiamen University, Xiamen, Fujian, China.

ABSTRACT

Background: Alternative polyadenylation as a mechanism in gene expression regulation has been widely recognized in recent years. Arabidopsis polyadenylation factor PCFS4 was shown to function in leaf development and in flowering time control. The function of PCFS4 in controlling flowering time was correlated with the alternative polyadenylation of FCA, a flowering time regulator. However, genetic evidence suggested additional targets of PCFS4 that may mediate its function in both flowering time and leaf development.

Methodology/principal findings: To identify further targets, we investigated the whole transcriptome of a PCFS4 mutant using Affymetrix Arabidopsis genomic tiling 1.0R array and developed a data analysis pipeline, termed RADPRE (Ratio-based Analysis of Differential mRNA Processing and Expression). In RADPRE, ratios of normalized probe intensities between wild type Columbia and a pcfs4 mutant were first generated. By doing so, one of the major problems of tiling array data--variations caused by differential probe affinity--was significantly alleviated. With the probe ratios as inputs, a hierarchy of statistical tests was carried out to identify differentially processed genes (DPG) and differentially expressed genes (DEG). The false discovery rate (FDR) of this analysis was estimated by using the balanced random combinations of Col/pcfs4 and pcfs4/Col ratios as inputs. Gene Ontology (GO) analysis of the DPGs and DEGs revealed potential new roles of PCFS4 in stress responses besides flowering time regulation.

Conclusion/significance: We identified 68 DPGs and 114 DEGs with FDR at 1% and 2%, respectively. Most of the 68 DPGs were subjected to alternative polyadenylation, splicing or transcription initiation. Quantitative PCR analysis of a set of DPGs confirmed that most of these genes were truly differentially processed in pcfs4 mutant plants. The enriched GO term "regulation of flower development" among PCFS4 targets further indicated the efficacy of the RADPRE pipeline. This simple but effective program is available upon request.

Show MeSH
Schematics of how a DPG gene could be identified based on the ratios of its exons between wild type Col (WT) and the pcfs4 mutant (MU).The gene structure was shown on the top of the graph with filled boxes denoting exons, lines denoting introns, and the short lines under the exons denoting the tiling array probes. The gene could generate two transcripts, a long (LT) and a short (ST), with LT derived from a distal poly(A) site and containing exon 3 (hatched box), and ST from a proximal poly(A) site within intron 2. The thickness of the box represents the relative abundance of the transcripts. The relative abundance of two transcripts was altered between WT and MU due to the shift of the poly(A) site usage between WT and MU. The measured abundance of each exon was based on its corresponding probes, which reflected the sum of two transcripts. The measured abundance of exon 1 and 2 was the same between WT and MU, but that of exon 3 was different. Therefore, the ratio of exon 1 and 2 between WT and MU was equal to 1, but different from the ratio of exon 3. In that case, the poly(A) site choice was not affected by the mutant and the relative abundance of the two transcripts would be the same between WT and MU. The ratio of the measured abundance between WT and MU would be equal for all three exons.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3045369&req=5

pone-0014719-g005: Schematics of how a DPG gene could be identified based on the ratios of its exons between wild type Col (WT) and the pcfs4 mutant (MU).The gene structure was shown on the top of the graph with filled boxes denoting exons, lines denoting introns, and the short lines under the exons denoting the tiling array probes. The gene could generate two transcripts, a long (LT) and a short (ST), with LT derived from a distal poly(A) site and containing exon 3 (hatched box), and ST from a proximal poly(A) site within intron 2. The thickness of the box represents the relative abundance of the transcripts. The relative abundance of two transcripts was altered between WT and MU due to the shift of the poly(A) site usage between WT and MU. The measured abundance of each exon was based on its corresponding probes, which reflected the sum of two transcripts. The measured abundance of exon 1 and 2 was the same between WT and MU, but that of exon 3 was different. Therefore, the ratio of exon 1 and 2 between WT and MU was equal to 1, but different from the ratio of exon 3. In that case, the poly(A) site choice was not affected by the mutant and the relative abundance of the two transcripts would be the same between WT and MU. The ratio of the measured abundance between WT and MU would be equal for all three exons.

Mentions: The underlying concept to identify the DPG targets of PCFS4 is depicted in Figure 5. With a given unknown gene, it was assumed the gene had two major transcripts derived from its pre-mRNA by APA, as in the case of FCA [13]. For example, if the choice of poly(A) sites is controlled by PCFS4, then the ratio of exon 3 abundance between WT and pcfs4 would be altered such that the ratio of exon 3 would deviate from 1 (more or less than 1) while the ratios of exon 1 and exon 2 should be equal to 1 (Figure 5). This case could be generalized as follows: if the APA of a given gene is controlled by PCFS4, then 1) the WT/mutant ratio of one or more exons would not be equal to one, and 2) the ratio of at least one exon would not be equal to the ratios of other exons. This generalization would also hold true if a gene's splicing and/or transcription initiation was affected by PCFS4. In the case of a given gene being a DEG target, which was defined as whole gene expression difference rather than an individual exon difference, the ratios between WT and pcfs4 of all exons for this gene would have the same degree of deviation from 1 (if larger than 1, under expressed in pcfs4; if smaller than 1, over expressed in pcfs4).


Ratio-based analysis of differential mRNA processing and expression of a polyadenylation factor mutant pcfs4 using arabidopsis tiling microarray.

Zheng J, Xing D, Wu X, Shen Y, Kroll DM, Ji G, Li QQ - PLoS ONE (2011)

Schematics of how a DPG gene could be identified based on the ratios of its exons between wild type Col (WT) and the pcfs4 mutant (MU).The gene structure was shown on the top of the graph with filled boxes denoting exons, lines denoting introns, and the short lines under the exons denoting the tiling array probes. The gene could generate two transcripts, a long (LT) and a short (ST), with LT derived from a distal poly(A) site and containing exon 3 (hatched box), and ST from a proximal poly(A) site within intron 2. The thickness of the box represents the relative abundance of the transcripts. The relative abundance of two transcripts was altered between WT and MU due to the shift of the poly(A) site usage between WT and MU. The measured abundance of each exon was based on its corresponding probes, which reflected the sum of two transcripts. The measured abundance of exon 1 and 2 was the same between WT and MU, but that of exon 3 was different. Therefore, the ratio of exon 1 and 2 between WT and MU was equal to 1, but different from the ratio of exon 3. In that case, the poly(A) site choice was not affected by the mutant and the relative abundance of the two transcripts would be the same between WT and MU. The ratio of the measured abundance between WT and MU would be equal for all three exons.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3045369&req=5

pone-0014719-g005: Schematics of how a DPG gene could be identified based on the ratios of its exons between wild type Col (WT) and the pcfs4 mutant (MU).The gene structure was shown on the top of the graph with filled boxes denoting exons, lines denoting introns, and the short lines under the exons denoting the tiling array probes. The gene could generate two transcripts, a long (LT) and a short (ST), with LT derived from a distal poly(A) site and containing exon 3 (hatched box), and ST from a proximal poly(A) site within intron 2. The thickness of the box represents the relative abundance of the transcripts. The relative abundance of two transcripts was altered between WT and MU due to the shift of the poly(A) site usage between WT and MU. The measured abundance of each exon was based on its corresponding probes, which reflected the sum of two transcripts. The measured abundance of exon 1 and 2 was the same between WT and MU, but that of exon 3 was different. Therefore, the ratio of exon 1 and 2 between WT and MU was equal to 1, but different from the ratio of exon 3. In that case, the poly(A) site choice was not affected by the mutant and the relative abundance of the two transcripts would be the same between WT and MU. The ratio of the measured abundance between WT and MU would be equal for all three exons.
Mentions: The underlying concept to identify the DPG targets of PCFS4 is depicted in Figure 5. With a given unknown gene, it was assumed the gene had two major transcripts derived from its pre-mRNA by APA, as in the case of FCA [13]. For example, if the choice of poly(A) sites is controlled by PCFS4, then the ratio of exon 3 abundance between WT and pcfs4 would be altered such that the ratio of exon 3 would deviate from 1 (more or less than 1) while the ratios of exon 1 and exon 2 should be equal to 1 (Figure 5). This case could be generalized as follows: if the APA of a given gene is controlled by PCFS4, then 1) the WT/mutant ratio of one or more exons would not be equal to one, and 2) the ratio of at least one exon would not be equal to the ratios of other exons. This generalization would also hold true if a gene's splicing and/or transcription initiation was affected by PCFS4. In the case of a given gene being a DEG target, which was defined as whole gene expression difference rather than an individual exon difference, the ratios between WT and pcfs4 of all exons for this gene would have the same degree of deviation from 1 (if larger than 1, under expressed in pcfs4; if smaller than 1, over expressed in pcfs4).

Bottom Line: Quantitative PCR analysis of a set of DPGs confirmed that most of these genes were truly differentially processed in pcfs4 mutant plants.The enriched GO term "regulation of flower development" among PCFS4 targets further indicated the efficacy of the RADPRE pipeline.This simple but effective program is available upon request.

View Article: PubMed Central - PubMed

Affiliation: Department of Automation, Xiamen University, Xiamen, Fujian, China.

ABSTRACT

Background: Alternative polyadenylation as a mechanism in gene expression regulation has been widely recognized in recent years. Arabidopsis polyadenylation factor PCFS4 was shown to function in leaf development and in flowering time control. The function of PCFS4 in controlling flowering time was correlated with the alternative polyadenylation of FCA, a flowering time regulator. However, genetic evidence suggested additional targets of PCFS4 that may mediate its function in both flowering time and leaf development.

Methodology/principal findings: To identify further targets, we investigated the whole transcriptome of a PCFS4 mutant using Affymetrix Arabidopsis genomic tiling 1.0R array and developed a data analysis pipeline, termed RADPRE (Ratio-based Analysis of Differential mRNA Processing and Expression). In RADPRE, ratios of normalized probe intensities between wild type Columbia and a pcfs4 mutant were first generated. By doing so, one of the major problems of tiling array data--variations caused by differential probe affinity--was significantly alleviated. With the probe ratios as inputs, a hierarchy of statistical tests was carried out to identify differentially processed genes (DPG) and differentially expressed genes (DEG). The false discovery rate (FDR) of this analysis was estimated by using the balanced random combinations of Col/pcfs4 and pcfs4/Col ratios as inputs. Gene Ontology (GO) analysis of the DPGs and DEGs revealed potential new roles of PCFS4 in stress responses besides flowering time regulation.

Conclusion/significance: We identified 68 DPGs and 114 DEGs with FDR at 1% and 2%, respectively. Most of the 68 DPGs were subjected to alternative polyadenylation, splicing or transcription initiation. Quantitative PCR analysis of a set of DPGs confirmed that most of these genes were truly differentially processed in pcfs4 mutant plants. The enriched GO term "regulation of flower development" among PCFS4 targets further indicated the efficacy of the RADPRE pipeline. This simple but effective program is available upon request.

Show MeSH