Limits...
Multiway real-time PCR gene expression profiling in yeast Saccharomyces cerevisiae reveals altered transcriptional response of ADH-genes to glucose stimuli.

Ståhlberg A, Elbing K, Andrade-Garda JM, Sjögreen B, Forootan A, Kubista M - BMC Genomics (2008)

Bottom Line: In multiway studies samples are characterized by their expression profiles to monitor changes over time, effect of treatment, drug dosage etc.The data are analyzed by matrix-augmented PCA, which is a generalization of PCA for 3-way data, and the results are confirmed by hierarchical clustering and clustering by Kohonen self-organizing map.The technique also identifies genes that show perturbed expression in specific strains.

View Article: PubMed Central - HTML - PubMed

Affiliation: TATAA Biocenter, Odinsgatan 28, 411 03 Göteborg, Sweden. anders.stahlberg@neuro.gu.se

ABSTRACT

Background: The large sensitivity, high reproducibility and essentially unlimited dynamic range of real-time PCR to measure gene expression in complex samples provides the opportunity for powerful multivariate and multiway studies of biological phenomena. In multiway studies samples are characterized by their expression profiles to monitor changes over time, effect of treatment, drug dosage etc. Here we perform a multiway study of the temporal response of four yeast Saccharomyces cerevisiae strains with different glucose uptake rates upon altered metabolic conditions.

Results: We measured the expression of 18 genes as function of time after addition of glucose to four strains of yeast grown in ethanol. The data are analyzed by matrix-augmented PCA, which is a generalization of PCA for 3-way data, and the results are confirmed by hierarchical clustering and clustering by Kohonen self-organizing map. Our approach identifies gene groups that respond similarly to the change of nutrient, and genes that behave differently in mutant strains. Of particular interest is our finding that ADH4 and ADH6 show a behavior typical of glucose-induced genes, while ADH3 and ADH5 are repressed after glucose addition.

Conclusion: Multiway real-time PCR gene expression profiling is a powerful technique which can be utilized to characterize functions of new genes by, for example, comparing their temporal response after perturbation in different genetic variants of the studied subject. The technique also identifies genes that show perturbed expression in specific strains.

Show MeSH

Related in: MedlinePlus

Kohonen self-organizing maps for the augmented wild-type, HXT-HXT7 and HXT-TM6* yeast. A map with six cells was used to force classification into six groups. A learning rate of 0.10, 2 neighbors and 10000 iterations were used. This gave the following groups: (1, 1) FBP1, ADH2, MDH2 (all strains), SUC2 (HXT-HXT7) HSP12 (HXT-HXT7 and HXT-TM6*), ADH3 (wild-type and HXT-HXT7) and ADH5 (wild-type); (1, 2) SUC2 (HXT-TM6*), HSP12 (wild-type) and ADH4 (HXT-HXT7); (1, 3) ADH1, PDC1, TPI1 (all strains), MIG1 (wild-type and HXT-HXT7), PGK1 (wild-type), ADH4 (HXT-TM6*) (2, 1) CYC1 (all strains), SUC2 (wild-type), ADH5 (HXT-HXT7 and HXT-TM6*), ADH3 (HXT-TM6*); (2, 2) PGK1 (HXT-TM6*); (2, 3) MIG1 (HXT-HXT7), PGK1 (HXT-HXT7), ADH4 (wild-type), ADH6 (all strains). The following colors and symbols are used: Glucose-induced genes (blue), glucose-repressed genes (red), ADH3-6 (yellow), HSP12 (black), CYC1 (green), wild-type (circles), HXT-HXT7 (squares) and HXT-TM6* (triangles).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2335116&req=5

Figure 7: Kohonen self-organizing maps for the augmented wild-type, HXT-HXT7 and HXT-TM6* yeast. A map with six cells was used to force classification into six groups. A learning rate of 0.10, 2 neighbors and 10000 iterations were used. This gave the following groups: (1, 1) FBP1, ADH2, MDH2 (all strains), SUC2 (HXT-HXT7) HSP12 (HXT-HXT7 and HXT-TM6*), ADH3 (wild-type and HXT-HXT7) and ADH5 (wild-type); (1, 2) SUC2 (HXT-TM6*), HSP12 (wild-type) and ADH4 (HXT-HXT7); (1, 3) ADH1, PDC1, TPI1 (all strains), MIG1 (wild-type and HXT-HXT7), PGK1 (wild-type), ADH4 (HXT-TM6*) (2, 1) CYC1 (all strains), SUC2 (wild-type), ADH5 (HXT-HXT7 and HXT-TM6*), ADH3 (HXT-TM6*); (2, 2) PGK1 (HXT-TM6*); (2, 3) MIG1 (HXT-HXT7), PGK1 (HXT-HXT7), ADH4 (wild-type), ADH6 (all strains). The following colors and symbols are used: Glucose-induced genes (blue), glucose-repressed genes (red), ADH3-6 (yellow), HSP12 (black), CYC1 (green), wild-type (circles), HXT-HXT7 (squares) and HXT-TM6* (triangles).

Mentions: There are two strategies to design SOM. A fairly large SOM, typically of dimension n × n, with n being the number of genes, can be used. In such SOM genes are sparse and only rarely is more than one gene found in one neuron. There are no clear boundaries between groups, because the SOM surface changes irregularly and distances between the genes are not proportional to the differences between them. Still, genes with similar expressions will be close to each other in the SOM, but there will be no space between groups of genes as in the case of PCA. This makes it hard to assign new groups and large SOMs are therefore more suitable to validate previous classifications based on PCA and hierarchical clustering. The alternative is to use a small SOM, with a rather small number of neurons. This forces the genes to cluster in the few neurons available, thereby creating groups. Testing different parameters for the generation of SOMs and classification of the autoscaled catenated wild-type, HXT-HXT7 and HXT-TM6* data, we found that a 3 × 2 SOM gave highly reproducible groupings, independent of the learning rate and the number of neighbors. Figure 7 shows the groups formed in such SOM based on a learning rate of 0.10, 2 neighbors and 10000 iterations. The groups agree very well with the PCA classification. In cell (2, 1) we find CYC1, ADH5 from HXT-TM6* and HXT-HXT7, ADH3 from HXT-TM6* and wild-type SUC2. These are the genes found in the bottom left of the PC1 vs. PC2 loadings scatter plot (Figure 5). Cell (2, 2) contains only PGK1 in HXT-TM6*, which was concluded above to have aberrant expression profile both by PCA and hierarchical clustering. In cell (1, 3) we find the glucose induced genes and ADH4 in HXT-TM6* that in the loadings scatter plot are found at negative PC1 and a PC2 around zero. Cell (1, 1) contains most of the glucose-repressed genes, HSP12 in HXT-TM6* and HXT-HXT7, ADH3 in HXT7 and wild-type, and ADH5 in wild-type, which in the PCA loadings scatter plot are found at positive PC1, and slightly positive PC2. Cell (1, 2) contains HSP12 from wild-type, SUC2 from HXT-TM6* and ADH4 from HXT-HXT7. In the PCA loadings scatter plot, HXT-TM6* SUC2 and wild-type HSP12 are also close to each other at positive PC2 and a PC1 around zero. ADH4 in HXT-HXT7 is also in this region, although it is closer to another group of genes that in the SOM is found in unit (2, 3). These are ADH6 in all three strains, ADH4 in wild-type, and MIG1 and PGK1 in HXT-HXT7. The same genes are found in the top left corner of the PC1 vs. PC2 loadings scatter plot.


Multiway real-time PCR gene expression profiling in yeast Saccharomyces cerevisiae reveals altered transcriptional response of ADH-genes to glucose stimuli.

Ståhlberg A, Elbing K, Andrade-Garda JM, Sjögreen B, Forootan A, Kubista M - BMC Genomics (2008)

Kohonen self-organizing maps for the augmented wild-type, HXT-HXT7 and HXT-TM6* yeast. A map with six cells was used to force classification into six groups. A learning rate of 0.10, 2 neighbors and 10000 iterations were used. This gave the following groups: (1, 1) FBP1, ADH2, MDH2 (all strains), SUC2 (HXT-HXT7) HSP12 (HXT-HXT7 and HXT-TM6*), ADH3 (wild-type and HXT-HXT7) and ADH5 (wild-type); (1, 2) SUC2 (HXT-TM6*), HSP12 (wild-type) and ADH4 (HXT-HXT7); (1, 3) ADH1, PDC1, TPI1 (all strains), MIG1 (wild-type and HXT-HXT7), PGK1 (wild-type), ADH4 (HXT-TM6*) (2, 1) CYC1 (all strains), SUC2 (wild-type), ADH5 (HXT-HXT7 and HXT-TM6*), ADH3 (HXT-TM6*); (2, 2) PGK1 (HXT-TM6*); (2, 3) MIG1 (HXT-HXT7), PGK1 (HXT-HXT7), ADH4 (wild-type), ADH6 (all strains). The following colors and symbols are used: Glucose-induced genes (blue), glucose-repressed genes (red), ADH3-6 (yellow), HSP12 (black), CYC1 (green), wild-type (circles), HXT-HXT7 (squares) and HXT-TM6* (triangles).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2335116&req=5

Figure 7: Kohonen self-organizing maps for the augmented wild-type, HXT-HXT7 and HXT-TM6* yeast. A map with six cells was used to force classification into six groups. A learning rate of 0.10, 2 neighbors and 10000 iterations were used. This gave the following groups: (1, 1) FBP1, ADH2, MDH2 (all strains), SUC2 (HXT-HXT7) HSP12 (HXT-HXT7 and HXT-TM6*), ADH3 (wild-type and HXT-HXT7) and ADH5 (wild-type); (1, 2) SUC2 (HXT-TM6*), HSP12 (wild-type) and ADH4 (HXT-HXT7); (1, 3) ADH1, PDC1, TPI1 (all strains), MIG1 (wild-type and HXT-HXT7), PGK1 (wild-type), ADH4 (HXT-TM6*) (2, 1) CYC1 (all strains), SUC2 (wild-type), ADH5 (HXT-HXT7 and HXT-TM6*), ADH3 (HXT-TM6*); (2, 2) PGK1 (HXT-TM6*); (2, 3) MIG1 (HXT-HXT7), PGK1 (HXT-HXT7), ADH4 (wild-type), ADH6 (all strains). The following colors and symbols are used: Glucose-induced genes (blue), glucose-repressed genes (red), ADH3-6 (yellow), HSP12 (black), CYC1 (green), wild-type (circles), HXT-HXT7 (squares) and HXT-TM6* (triangles).
Mentions: There are two strategies to design SOM. A fairly large SOM, typically of dimension n × n, with n being the number of genes, can be used. In such SOM genes are sparse and only rarely is more than one gene found in one neuron. There are no clear boundaries between groups, because the SOM surface changes irregularly and distances between the genes are not proportional to the differences between them. Still, genes with similar expressions will be close to each other in the SOM, but there will be no space between groups of genes as in the case of PCA. This makes it hard to assign new groups and large SOMs are therefore more suitable to validate previous classifications based on PCA and hierarchical clustering. The alternative is to use a small SOM, with a rather small number of neurons. This forces the genes to cluster in the few neurons available, thereby creating groups. Testing different parameters for the generation of SOMs and classification of the autoscaled catenated wild-type, HXT-HXT7 and HXT-TM6* data, we found that a 3 × 2 SOM gave highly reproducible groupings, independent of the learning rate and the number of neighbors. Figure 7 shows the groups formed in such SOM based on a learning rate of 0.10, 2 neighbors and 10000 iterations. The groups agree very well with the PCA classification. In cell (2, 1) we find CYC1, ADH5 from HXT-TM6* and HXT-HXT7, ADH3 from HXT-TM6* and wild-type SUC2. These are the genes found in the bottom left of the PC1 vs. PC2 loadings scatter plot (Figure 5). Cell (2, 2) contains only PGK1 in HXT-TM6*, which was concluded above to have aberrant expression profile both by PCA and hierarchical clustering. In cell (1, 3) we find the glucose induced genes and ADH4 in HXT-TM6* that in the loadings scatter plot are found at negative PC1 and a PC2 around zero. Cell (1, 1) contains most of the glucose-repressed genes, HSP12 in HXT-TM6* and HXT-HXT7, ADH3 in HXT7 and wild-type, and ADH5 in wild-type, which in the PCA loadings scatter plot are found at positive PC1, and slightly positive PC2. Cell (1, 2) contains HSP12 from wild-type, SUC2 from HXT-TM6* and ADH4 from HXT-HXT7. In the PCA loadings scatter plot, HXT-TM6* SUC2 and wild-type HSP12 are also close to each other at positive PC2 and a PC1 around zero. ADH4 in HXT-HXT7 is also in this region, although it is closer to another group of genes that in the SOM is found in unit (2, 3). These are ADH6 in all three strains, ADH4 in wild-type, and MIG1 and PGK1 in HXT-HXT7. The same genes are found in the top left corner of the PC1 vs. PC2 loadings scatter plot.

Bottom Line: In multiway studies samples are characterized by their expression profiles to monitor changes over time, effect of treatment, drug dosage etc.The data are analyzed by matrix-augmented PCA, which is a generalization of PCA for 3-way data, and the results are confirmed by hierarchical clustering and clustering by Kohonen self-organizing map.The technique also identifies genes that show perturbed expression in specific strains.

View Article: PubMed Central - HTML - PubMed

Affiliation: TATAA Biocenter, Odinsgatan 28, 411 03 Göteborg, Sweden. anders.stahlberg@neuro.gu.se

ABSTRACT

Background: The large sensitivity, high reproducibility and essentially unlimited dynamic range of real-time PCR to measure gene expression in complex samples provides the opportunity for powerful multivariate and multiway studies of biological phenomena. In multiway studies samples are characterized by their expression profiles to monitor changes over time, effect of treatment, drug dosage etc. Here we perform a multiway study of the temporal response of four yeast Saccharomyces cerevisiae strains with different glucose uptake rates upon altered metabolic conditions.

Results: We measured the expression of 18 genes as function of time after addition of glucose to four strains of yeast grown in ethanol. The data are analyzed by matrix-augmented PCA, which is a generalization of PCA for 3-way data, and the results are confirmed by hierarchical clustering and clustering by Kohonen self-organizing map. Our approach identifies gene groups that respond similarly to the change of nutrient, and genes that behave differently in mutant strains. Of particular interest is our finding that ADH4 and ADH6 show a behavior typical of glucose-induced genes, while ADH3 and ADH5 are repressed after glucose addition.

Conclusion: Multiway real-time PCR gene expression profiling is a powerful technique which can be utilized to characterize functions of new genes by, for example, comparing their temporal response after perturbation in different genetic variants of the studied subject. The technique also identifies genes that show perturbed expression in specific strains.

Show MeSH
Related in: MedlinePlus