Limits...
Improved coverage of cDNA-AFLP by sequential digestion of immobilized cDNA.

Weiberg A, Pöhler D, Morgenstern B, Karlovsky P - BMC Genomics (2008)

Bottom Line: cDNA-AFLP is a transcriptomics technique which does not require prior sequence information and can therefore be used as a gene discovery tool.Some transcripts are represented by more than one fragment while other escape detection, causing redundancy and reducing the coverage of the analysis, respectively.For A. thaliana, human and mice transcriptome, the use of two marking enzymes and three sequentially applied releasing enzymes for each of the marking enzymes is recommended.

View Article: PubMed Central - HTML - PubMed

Affiliation: Molecular Phytopathology and Mycotoxin Research Division, University of Goettingen, Goettingen, Germany. aweiber1@gwdg.de

ABSTRACT

Background: cDNA-AFLP is a transcriptomics technique which does not require prior sequence information and can therefore be used as a gene discovery tool. The method is based on selective amplification of cDNA fragments generated by restriction endonucleases, electrophoretic separation of the products and comparison of the band patterns between treated samples and controls. Unequal distribution of restriction sites used to generate cDNA fragments negatively affects the performance of cDNA-AFLP. Some transcripts are represented by more than one fragment while other escape detection, causing redundancy and reducing the coverage of the analysis, respectively.

Results: With the goal of improving the coverage of cDNA-AFLP without increasing its redundancy, we designed a modified cDNA-AFLP protocol. Immobilized cDNA is sequentially digested with several restriction endonucleases and the released DNA fragments are collected in mutually exclusive pools. To investigate the performance of the protocol, software tool MECS (Multiple Enzyme cDNA-AFLP Simulation) was written in Perl. cDNA-AFLP protocols described in the literature and the new sequential digestion protocol were simulated on sets of cDNA sequences from mouse, human and Arabidopsis thaliana. The redundancy and coverage, the total number of PCR reactions, and the average fragment length were calculated for each protocol and cDNA set.

Conclusion: Simulation revealed that sequential digestion of immobilized cDNA followed by the partitioning of released fragments into mutually exclusive pools outperformed other cDNA-AFLP protocols in terms of coverage, redundancy, fragment length, and the total number of PCRs. Primers generating 30 to 70 amplicons per PCR provided the highest fraction of electrophoretically distinguishable fragments suitable for normalization. For A. thaliana, human and mice transcriptome, the use of two marking enzymes and three sequentially applied releasing enzymes for each of the marking enzymes is recommended.

Show MeSH

Related in: MedlinePlus

Effect of the number of PCR products on the fraction of analyzable fragments. For RefSeq EST data sets (Tab. 1), randomly selected enzyme combinations were used to simulate cDNA-AFLP with the sequential digestion. One to three selective nucleotides were attached to PCR primers. For each PCR reaction, the fraction of analyzable fragments (fragments between 40 bp and 700 bp with a length that occurred only once, at least 20 products per PCR) was plotted against the number of PCR products. Vertical bars indicate standard error. Black: Arabidopsis; red: mouse; green: human.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2577664&req=5

Figure 3: Effect of the number of PCR products on the fraction of analyzable fragments. For RefSeq EST data sets (Tab. 1), randomly selected enzyme combinations were used to simulate cDNA-AFLP with the sequential digestion. One to three selective nucleotides were attached to PCR primers. For each PCR reaction, the fraction of analyzable fragments (fragments between 40 bp and 700 bp with a length that occurred only once, at least 20 products per PCR) was plotted against the number of PCR products. Vertical bars indicate standard error. Black: Arabidopsis; red: mouse; green: human.

Mentions: To determine the optimal number of selective nucleotides N, we need to know the size of the fragment pool that will be partitioned by PCR and the optimal number of products per PCR reaction. The probability of co-migration during electrophoresis increases with decreasing N, while the probability that a PCR reaction will generate fewer than 20 products grows with N. To determine the optimal number of PCR reactions, we simulated cDNA-AFLP with the sequential digestion protocol for two marking and two to three releasing enzymes. Enzyme combinations were randomly selected from Tab. 2. PCR was simulated with primers containing one to three selective nucleotides (total for both primers). For each PCR reaction, the products were sorted by size, and fragments of any size that occurred more than once were eliminated. Furthermore, products shorter than 40 bp or larger than 700 bp were eliminated. The fraction of fragments remaining in the set after the treatment was scored "analyzable." The results of these simulations are summarized in Fig. 3. Even with the optimal number of PCR products, only about 75% of the fragments are analyzable. Therefore, coverage values predicted by simulations that do not take co-migration into account and do not eliminate fragments that are too short or too long have to be reduced accordingly.


Improved coverage of cDNA-AFLP by sequential digestion of immobilized cDNA.

Weiberg A, Pöhler D, Morgenstern B, Karlovsky P - BMC Genomics (2008)

Effect of the number of PCR products on the fraction of analyzable fragments. For RefSeq EST data sets (Tab. 1), randomly selected enzyme combinations were used to simulate cDNA-AFLP with the sequential digestion. One to three selective nucleotides were attached to PCR primers. For each PCR reaction, the fraction of analyzable fragments (fragments between 40 bp and 700 bp with a length that occurred only once, at least 20 products per PCR) was plotted against the number of PCR products. Vertical bars indicate standard error. Black: Arabidopsis; red: mouse; green: human.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2577664&req=5

Figure 3: Effect of the number of PCR products on the fraction of analyzable fragments. For RefSeq EST data sets (Tab. 1), randomly selected enzyme combinations were used to simulate cDNA-AFLP with the sequential digestion. One to three selective nucleotides were attached to PCR primers. For each PCR reaction, the fraction of analyzable fragments (fragments between 40 bp and 700 bp with a length that occurred only once, at least 20 products per PCR) was plotted against the number of PCR products. Vertical bars indicate standard error. Black: Arabidopsis; red: mouse; green: human.
Mentions: To determine the optimal number of selective nucleotides N, we need to know the size of the fragment pool that will be partitioned by PCR and the optimal number of products per PCR reaction. The probability of co-migration during electrophoresis increases with decreasing N, while the probability that a PCR reaction will generate fewer than 20 products grows with N. To determine the optimal number of PCR reactions, we simulated cDNA-AFLP with the sequential digestion protocol for two marking and two to three releasing enzymes. Enzyme combinations were randomly selected from Tab. 2. PCR was simulated with primers containing one to three selective nucleotides (total for both primers). For each PCR reaction, the products were sorted by size, and fragments of any size that occurred more than once were eliminated. Furthermore, products shorter than 40 bp or larger than 700 bp were eliminated. The fraction of fragments remaining in the set after the treatment was scored "analyzable." The results of these simulations are summarized in Fig. 3. Even with the optimal number of PCR products, only about 75% of the fragments are analyzable. Therefore, coverage values predicted by simulations that do not take co-migration into account and do not eliminate fragments that are too short or too long have to be reduced accordingly.

Bottom Line: cDNA-AFLP is a transcriptomics technique which does not require prior sequence information and can therefore be used as a gene discovery tool.Some transcripts are represented by more than one fragment while other escape detection, causing redundancy and reducing the coverage of the analysis, respectively.For A. thaliana, human and mice transcriptome, the use of two marking enzymes and three sequentially applied releasing enzymes for each of the marking enzymes is recommended.

View Article: PubMed Central - HTML - PubMed

Affiliation: Molecular Phytopathology and Mycotoxin Research Division, University of Goettingen, Goettingen, Germany. aweiber1@gwdg.de

ABSTRACT

Background: cDNA-AFLP is a transcriptomics technique which does not require prior sequence information and can therefore be used as a gene discovery tool. The method is based on selective amplification of cDNA fragments generated by restriction endonucleases, electrophoretic separation of the products and comparison of the band patterns between treated samples and controls. Unequal distribution of restriction sites used to generate cDNA fragments negatively affects the performance of cDNA-AFLP. Some transcripts are represented by more than one fragment while other escape detection, causing redundancy and reducing the coverage of the analysis, respectively.

Results: With the goal of improving the coverage of cDNA-AFLP without increasing its redundancy, we designed a modified cDNA-AFLP protocol. Immobilized cDNA is sequentially digested with several restriction endonucleases and the released DNA fragments are collected in mutually exclusive pools. To investigate the performance of the protocol, software tool MECS (Multiple Enzyme cDNA-AFLP Simulation) was written in Perl. cDNA-AFLP protocols described in the literature and the new sequential digestion protocol were simulated on sets of cDNA sequences from mouse, human and Arabidopsis thaliana. The redundancy and coverage, the total number of PCR reactions, and the average fragment length were calculated for each protocol and cDNA set.

Conclusion: Simulation revealed that sequential digestion of immobilized cDNA followed by the partitioning of released fragments into mutually exclusive pools outperformed other cDNA-AFLP protocols in terms of coverage, redundancy, fragment length, and the total number of PCRs. Primers generating 30 to 70 amplicons per PCR provided the highest fraction of electrophoretically distinguishable fragments suitable for normalization. For A. thaliana, human and mice transcriptome, the use of two marking enzymes and three sequentially applied releasing enzymes for each of the marking enzymes is recommended.

Show MeSH
Related in: MedlinePlus