Limits...
Single-species microarrays and comparative transcriptomics.

Chain FJ, Ilieva D, Evans BJ - PLoS ONE (2008)

Bottom Line: Prefabricated expression microarrays are currently available for only a few species but methods have been proposed to extend their application to comparisons between divergent genomes.Here we demonstrate that the hybridization intensity of genomic DNA is a poor basis on which to select unbiased probes on Affymetrix expression arrays for studies of comparative transcriptomics, and that doing so produces spurious results.These findings also highlight important limitations of single-species microarrays for studies of comparative transcriptomics of polyploid species.

View Article: PubMed Central - PubMed

Affiliation: Department of Biology, Center for Environmental Genomics, McMaster University, Hamilton, Ontario, Canada.

ABSTRACT

Background: Prefabricated expression microarrays are currently available for only a few species but methods have been proposed to extend their application to comparisons between divergent genomes.

Methodology/principal findings: Here we demonstrate that the hybridization intensity of genomic DNA is a poor basis on which to select unbiased probes on Affymetrix expression arrays for studies of comparative transcriptomics, and that doing so produces spurious results. We used the Affymetrix Xenopus laevis microarray to evaluate expression divergence between X. laevis, X. borealis, and their F1 hybrids. When data are analyzed with probes that interrogate only sequences with confirmed identity in both species, we recover results that differ substantially analyses that use genomic DNA hybridizations to select probes.

Conclusions/significance: Our findings have implications for the experimental design of comparative expression studies that use single-species microarrays, and for our understanding of divergent expression in hybrid clawed frogs. These findings also highlight important limitations of single-species microarrays for studies of comparative transcriptomics of polyploid species.

Show MeSH
Genomic hybridization intensities (gDNA intensity) of XL, XB, and XM vary with respect to the non-target to target ratio of these intensities (gDNA ratio).This graph depicts the median gDNA intensities of all probes on the chip ranked by their gDNA ratio into 25 bins; each bin contains 10,000 probes except the 25th bin, which contains 7852 probes. The area in gray corresponds with the range of gDNA ratios of probes that are retained using the method of Malone et al. (2007). XL gDNA ratios are represented by filled symbols and non-target gDNA ratios are represented by unfilled symbols. Shown are relationships between the median gDNA intensity of each bin and the median gDNA ratio of each bin for (A) our XM and XL gDNA hybridizations, (B) the XM and XL gDNA hybridizations of Malone et al. (2007), and (C) our XB and XL gDNA hybridizations.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2533705&req=5

pone-0003279-g001: Genomic hybridization intensities (gDNA intensity) of XL, XB, and XM vary with respect to the non-target to target ratio of these intensities (gDNA ratio).This graph depicts the median gDNA intensities of all probes on the chip ranked by their gDNA ratio into 25 bins; each bin contains 10,000 probes except the 25th bin, which contains 7852 probes. The area in gray corresponds with the range of gDNA ratios of probes that are retained using the method of Malone et al. (2007). XL gDNA ratios are represented by filled symbols and non-target gDNA ratios are represented by unfilled symbols. Shown are relationships between the median gDNA intensity of each bin and the median gDNA ratio of each bin for (A) our XM and XL gDNA hybridizations, (B) the XM and XL gDNA hybridizations of Malone et al. (2007), and (C) our XB and XL gDNA hybridizations.

Mentions: We compared our gDNA hybridizations to those of [17], [18]. We ranked all of the probes on the chip by the gDNA hybridization intensity and then divided these ranks into 25 bins. Comparison to the gDNA ratio of each probe indicates that the median intensity of hybridization was lower in the non-target species (XM or XB) than the target species (XL) for most bins (Fig. 1). Probes with a gDNA ratio near one tended to have lower gDNA hybridization intensities in both the non-target and the target species than other probes on the chip, and the target species (XL) tends to have a more dynamic relationship between probe intensity and the gDNA ratio. Thus, at least on the Affymetrix GeneChip® Xenopus laevis Genome Array, probe selection on the basis of a gDNA hybridization ratio near one appears to have an unintended consequence of retaining probes with low gDNA hybridization intensities in both species. This was true in gDNA hybridizations performed by our lab and also by another lab (Fig. 1), thus it is not attributable to differences in laboratory procedure.


Single-species microarrays and comparative transcriptomics.

Chain FJ, Ilieva D, Evans BJ - PLoS ONE (2008)

Genomic hybridization intensities (gDNA intensity) of XL, XB, and XM vary with respect to the non-target to target ratio of these intensities (gDNA ratio).This graph depicts the median gDNA intensities of all probes on the chip ranked by their gDNA ratio into 25 bins; each bin contains 10,000 probes except the 25th bin, which contains 7852 probes. The area in gray corresponds with the range of gDNA ratios of probes that are retained using the method of Malone et al. (2007). XL gDNA ratios are represented by filled symbols and non-target gDNA ratios are represented by unfilled symbols. Shown are relationships between the median gDNA intensity of each bin and the median gDNA ratio of each bin for (A) our XM and XL gDNA hybridizations, (B) the XM and XL gDNA hybridizations of Malone et al. (2007), and (C) our XB and XL gDNA hybridizations.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2533705&req=5

pone-0003279-g001: Genomic hybridization intensities (gDNA intensity) of XL, XB, and XM vary with respect to the non-target to target ratio of these intensities (gDNA ratio).This graph depicts the median gDNA intensities of all probes on the chip ranked by their gDNA ratio into 25 bins; each bin contains 10,000 probes except the 25th bin, which contains 7852 probes. The area in gray corresponds with the range of gDNA ratios of probes that are retained using the method of Malone et al. (2007). XL gDNA ratios are represented by filled symbols and non-target gDNA ratios are represented by unfilled symbols. Shown are relationships between the median gDNA intensity of each bin and the median gDNA ratio of each bin for (A) our XM and XL gDNA hybridizations, (B) the XM and XL gDNA hybridizations of Malone et al. (2007), and (C) our XB and XL gDNA hybridizations.
Mentions: We compared our gDNA hybridizations to those of [17], [18]. We ranked all of the probes on the chip by the gDNA hybridization intensity and then divided these ranks into 25 bins. Comparison to the gDNA ratio of each probe indicates that the median intensity of hybridization was lower in the non-target species (XM or XB) than the target species (XL) for most bins (Fig. 1). Probes with a gDNA ratio near one tended to have lower gDNA hybridization intensities in both the non-target and the target species than other probes on the chip, and the target species (XL) tends to have a more dynamic relationship between probe intensity and the gDNA ratio. Thus, at least on the Affymetrix GeneChip® Xenopus laevis Genome Array, probe selection on the basis of a gDNA hybridization ratio near one appears to have an unintended consequence of retaining probes with low gDNA hybridization intensities in both species. This was true in gDNA hybridizations performed by our lab and also by another lab (Fig. 1), thus it is not attributable to differences in laboratory procedure.

Bottom Line: Prefabricated expression microarrays are currently available for only a few species but methods have been proposed to extend their application to comparisons between divergent genomes.Here we demonstrate that the hybridization intensity of genomic DNA is a poor basis on which to select unbiased probes on Affymetrix expression arrays for studies of comparative transcriptomics, and that doing so produces spurious results.These findings also highlight important limitations of single-species microarrays for studies of comparative transcriptomics of polyploid species.

View Article: PubMed Central - PubMed

Affiliation: Department of Biology, Center for Environmental Genomics, McMaster University, Hamilton, Ontario, Canada.

ABSTRACT

Background: Prefabricated expression microarrays are currently available for only a few species but methods have been proposed to extend their application to comparisons between divergent genomes.

Methodology/principal findings: Here we demonstrate that the hybridization intensity of genomic DNA is a poor basis on which to select unbiased probes on Affymetrix expression arrays for studies of comparative transcriptomics, and that doing so produces spurious results. We used the Affymetrix Xenopus laevis microarray to evaluate expression divergence between X. laevis, X. borealis, and their F1 hybrids. When data are analyzed with probes that interrogate only sequences with confirmed identity in both species, we recover results that differ substantially analyses that use genomic DNA hybridizations to select probes.

Conclusions/significance: Our findings have implications for the experimental design of comparative expression studies that use single-species microarrays, and for our understanding of divergent expression in hybrid clawed frogs. These findings also highlight important limitations of single-species microarrays for studies of comparative transcriptomics of polyploid species.

Show MeSH