Limits...
GOGOT: a method for the identification of differentially expressed fragments from cDNA-AFLP data.

Kadota K, Araki R, Nakai Y, Abe M - Algorithms Mol Biol (2007)

Bottom Line: The output of the analysis is a highly reduced list of differentially expressed TDFs.The validity of the automated ranking of TDFs by the special statistic was confirmed by the visual evaluation of a third party.The current algorithm may be applied to other electrophoretic data and temporal microarray data.

View Article: PubMed Central - HTML - PubMed

Affiliation: Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan. kadota@iu.a.u-tokyo.ac.jp

ABSTRACT

Background: One-dimensional (1-D) electrophoretic data obtained using the cDNA-AFLP method have attracted great interest for the identification of differentially expressed transcript-derived fragments (TDFs). However, high-throughput analysis of the cDNA-AFLP data is currently limited by the need for labor-intensive visual evaluation of multiple electropherograms. We would like to have high-throughput ways of identifying such TDFs.

Results: We describe a method, GOGOT, which automatically detects the differentially expressed TDFs in a set of time-course electropherograms. Analysis by GOGOT is conducted as follows: correction of fragment lengths of TDFs, alignment of identical TDFs across different electropherograms, normalization of peak heights, and identification of differentially expressed TDFs using a special statistic. The output of the analysis is a highly reduced list of differentially expressed TDFs. Visual evaluation confirmed that the peak alignment was performed perfectly for the TDFs by virtue of the correction of peak fragment lengths before alignment in step 1. The validity of the automated ranking of TDFs by the special statistic was confirmed by the visual evaluation of a third party.

Conclusion: GOGOT is useful for the automated detection of differentially expressed TDFs from cDNA-AFLP temporal electrophoretic data. The current algorithm may be applied to other electrophoretic data and temporal microarray data.

No MeSH data available.


Normalized peak fragment lengths in HiCEP electropherograms in Fig. 1. Note that individual TDFs are represented by tight clusters and all peaks in the cluster are of course correctly aligned. The alignment connected by black bold lines in Fig. 1a is represented by black dashed lines and sectioned when peak alignment is reapplied to the normalized electropherograms.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC1904450&req=5

Figure 2: Normalized peak fragment lengths in HiCEP electropherograms in Fig. 1. Note that individual TDFs are represented by tight clusters and all peaks in the cluster are of course correctly aligned. The alignment connected by black bold lines in Fig. 1a is represented by black dashed lines and sectioned when peak alignment is reapplied to the normalized electropherograms.

Mentions: In this paper, peak alignment for HiCEP electrophoretic data (Step 2) is performed using an algorithm based on complete linkage hierarchical clustering [20], though algorithms based on dynamic programming (DP) have widely been used for the purpose [21-25]. Perhaps a sophisticated DP-based method could perform accurate alignment such as shown in Fig. 2 for electropherograms in Fig. 1 without step 1. Nevertheless, the results of peak alignment for normalized electropherograms such as shown in Fig. 2 obtained from our two-step process (step 1 and 2) were satisfactory and those visual evaluations were very easy. The advantageous characteristics of our two-step approach over conventional DP-based methods [21-25] may be (i) easy visual evaluation by virtue of step 1 and (ii) easy traceability of why peaks are merged into a single TDF by virtue of a simple clustering-based method at step 2 (for details, see Methods). In general, labor-intensive visual evaluation of the electropherograms imposes bottlenecks on high-throughput expression analysis by electrophoretic methods including cDNA-AFLP [23]. Although there is currently no convincing rationale for choosing between the different methods, our two-step approach may eventually increase throughput. We demonstrate the feasibility of GOGOT in the rest of this section.


GOGOT: a method for the identification of differentially expressed fragments from cDNA-AFLP data.

Kadota K, Araki R, Nakai Y, Abe M - Algorithms Mol Biol (2007)

Normalized peak fragment lengths in HiCEP electropherograms in Fig. 1. Note that individual TDFs are represented by tight clusters and all peaks in the cluster are of course correctly aligned. The alignment connected by black bold lines in Fig. 1a is represented by black dashed lines and sectioned when peak alignment is reapplied to the normalized electropherograms.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC1904450&req=5

Figure 2: Normalized peak fragment lengths in HiCEP electropherograms in Fig. 1. Note that individual TDFs are represented by tight clusters and all peaks in the cluster are of course correctly aligned. The alignment connected by black bold lines in Fig. 1a is represented by black dashed lines and sectioned when peak alignment is reapplied to the normalized electropherograms.
Mentions: In this paper, peak alignment for HiCEP electrophoretic data (Step 2) is performed using an algorithm based on complete linkage hierarchical clustering [20], though algorithms based on dynamic programming (DP) have widely been used for the purpose [21-25]. Perhaps a sophisticated DP-based method could perform accurate alignment such as shown in Fig. 2 for electropherograms in Fig. 1 without step 1. Nevertheless, the results of peak alignment for normalized electropherograms such as shown in Fig. 2 obtained from our two-step process (step 1 and 2) were satisfactory and those visual evaluations were very easy. The advantageous characteristics of our two-step approach over conventional DP-based methods [21-25] may be (i) easy visual evaluation by virtue of step 1 and (ii) easy traceability of why peaks are merged into a single TDF by virtue of a simple clustering-based method at step 2 (for details, see Methods). In general, labor-intensive visual evaluation of the electropherograms imposes bottlenecks on high-throughput expression analysis by electrophoretic methods including cDNA-AFLP [23]. Although there is currently no convincing rationale for choosing between the different methods, our two-step approach may eventually increase throughput. We demonstrate the feasibility of GOGOT in the rest of this section.

Bottom Line: The output of the analysis is a highly reduced list of differentially expressed TDFs.The validity of the automated ranking of TDFs by the special statistic was confirmed by the visual evaluation of a third party.The current algorithm may be applied to other electrophoretic data and temporal microarray data.

View Article: PubMed Central - HTML - PubMed

Affiliation: Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan. kadota@iu.a.u-tokyo.ac.jp

ABSTRACT

Background: One-dimensional (1-D) electrophoretic data obtained using the cDNA-AFLP method have attracted great interest for the identification of differentially expressed transcript-derived fragments (TDFs). However, high-throughput analysis of the cDNA-AFLP data is currently limited by the need for labor-intensive visual evaluation of multiple electropherograms. We would like to have high-throughput ways of identifying such TDFs.

Results: We describe a method, GOGOT, which automatically detects the differentially expressed TDFs in a set of time-course electropherograms. Analysis by GOGOT is conducted as follows: correction of fragment lengths of TDFs, alignment of identical TDFs across different electropherograms, normalization of peak heights, and identification of differentially expressed TDFs using a special statistic. The output of the analysis is a highly reduced list of differentially expressed TDFs. Visual evaluation confirmed that the peak alignment was performed perfectly for the TDFs by virtue of the correction of peak fragment lengths before alignment in step 1. The validity of the automated ranking of TDFs by the special statistic was confirmed by the visual evaluation of a third party.

Conclusion: GOGOT is useful for the automated detection of differentially expressed TDFs from cDNA-AFLP temporal electrophoretic data. The current algorithm may be applied to other electrophoretic data and temporal microarray data.

No MeSH data available.