Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2003 Jul 10;4(1):27.
doi: 10.1186/1471-2164-4-27.

Whole-genome microarrays of fission yeast: characteristics, accuracy, reproducibility, and processing of array data

Rachel Lyne  1 , Gavin Burns, Juan Mata, Chris J Penkett, Gabriella Rustici, Dongrong Chen, Cordelia Langford, David Vetrie, Jürg Bähler
Affiliations

Whole-genome microarrays of fission yeast: characteristics, accuracy, reproducibility, and processing of array data

Rachel Lyne et al. BMC Genomics. .

Abstract

Background: The genome of the fission yeast Schizosaccharomyces pombe has recently been sequenced, setting the stage for the post-genomic era of this increasingly popular model organism. We have built fission yeast microarrays, optimised protocols to improve array performance, and carried out experiments to assess various characteristics of microarrays.

Results: We designed PCR primers to amplify specific probes (180-500 bp) for all known and predicted fission yeast genes, which are printed in duplicate onto separate regions of glass slides together with control elements (approximately 13,000 spots/slide). Fluorescence signal intensities depended on the size and intragenic position of the array elements, whereas the signal ratios were largely independent of element properties. Only the coding strand is covalently linked to the slides, and our array elements can discriminate transcriptional direction. The microarrays can distinguish sequences with up to 70% identity, above which cross-hybridisation contributes to the signal intensity. We tested the accuracy of signal ratios and measured the reproducibility of array data caused by biological and technical factors. Because the technical variability is lower, it is best to use samples prepared from independent biological experiments to obtain repeated measurements with swapping of fluorochromes to prevent dye bias. We also developed a script that discards unreliable data and performs a normalization to correct spatial artefacts.

Conclusions: This paper provides data for several microarray properties that are rarely measured. The results define critical parameters for microarray design and experiments and provide a framework to optimise and interpret array data. Our arrays give reproducible and accurate expression ratios with high sensitivity. The scripts for primer design and initial data processing as well as primer sequences and detailed protocols are available from our website.

PubMed Disclaimer

Figures

Figure 1
Figure 1
Effects of array element size and position on fluorescence signals and ratios. (A) PCR products of varying sizes (80–1500 bp) were used as array elements for two genes (mid1 and cdc12). In all cases, the 3' ends of the array elements were kept constant (~50 bp upstream of the stop codon). Top: fluorescence signals (local background subtracted) relative to array element size; the means and standard deviations of eight signal measurements are shown (two self-self experiments with two replicate measurements of both Cy3 and Cy5 each). Bottom: normalized ratios of signals (Cy5/Cy3) relative to array element size; the means and standard deviations of four measurements are shown (two self-self experiments with two replicate measurements each). (B) PCR products from varying positions within two genes (ags1 and tif471) were used as array elements. In all cases, the sizes of array elements were similar (~500 bp). Top: fluorescence signals (local background subtracted) relative to array element position (measured as distance of 3'ends of elements to stop codon); the means and standard deviations of eight signal measurements are shown (two self-self experiments with two replicate measurements of both Cy3 and Cy5 each). Bottom: normalized ratios of signals (Cy5/Cy3) relative to array element position; the means and standard deviations of four measurements are shown (two self-self experiments with two replicate measurements each).
Figure 2
Figure 2
Ratio-intensity plots before and after normalization. (A) log-transformed signal ratios (Cy3/Cy5) are plotted against the log-transformed products of signal intensities. Grey spots: discarded data that were filtered out by initial data processing (see text for details); blue spots: data used for further evaluation. Left side: temperature experiment with sample from cells grown at 25°C labelled with Cy5 and sample from cells grown at 30°C labelled with Cy3. Right side: identical sample labelled with Cy3 and Cy5 and hybridised on same array ('self-self' hybridisation): all signal ratios are expected to be 1, and the absence of differential gene expression is reflected by a tighter distribution of the spots. The number of the blue spots that were retained for data evaluation is 9161 (left) and 8560 (right). (B) As in (A) after normalization of the data using our local normalization scheme. (C) Overlay of spots before (blue) and after (red) normalization.
Figure 3
Figure 3
Correction of spatial artifacts by local normalization. (A) Distribution of signal ratios along the Y- (left) and X-axis (right) of the microarray slide before normalization. The data are from the same array as in Figure 2 (left side). Only spots giving usable data are shown. The groups of spots separated by small gaps reflect the 12 × 4 sub-grids of the array, each printed with a different spotting pin. (B) Distribution of signal ratios as in (A) after local normalization of the data. (C) Overlay of the data from (A) and (B).
Figure 4
Figure 4
Replicate data are more similar to each other after local normalization. The ratios of signal ratios from corresponding pairs of replicate spots on the same array were determined for all array elements (Ratio 1 / Ratio 2). Increased agreement between replicate measurements leads to ratios of signal ratios closer to 1. The data are from the same array as in Figure 2 (left side) and Figure 3. Left histogram: replicate data distribution before normalization; right histogram: replicate data distribution after normalization.
Figure 5
Figure 5
Linear readout range and detection limit of spiked RNA samples. 20 μg of S. pombe total RNA was spiked with five Bacillus subtilis mRNAs at various concentrations. Left side: lysA (3 pg), pheB (15 pg), thrB (30 pg), dapB (150 pg), and trpC (300 pg); right side: lysA (60 pg), pheB (300 pg), thrB (600 pg), and dapB (3000 pg). The median fluorescence intensities above local background of the B. subtilis control spots (determined from ~100 spots/transcript distributed evenly across the array) were plotted as a function of transcript concentration.
Figure 6
Figure 6
Unspecific hybridisation to similar array elements. PCR products from S. cerevisiae genes with various sequence similarities to S. pombe genes were used as array elements and hybridised with S. pombe samples. Hybridisation data are shown for 13 such genes showing 51%-79% identity to S. pombe genes across their entire lengths of ~200 bp. Blue bars: fluorescence signals (local background subtracted) picked up by the S. cerevisiae array elements; the means and standard deviations of four signal measurements are shown (one self-self experiment with two replicate measurements of both Cy3 and Cy5 each). Yellow bars: relative amount (in percentages) of 'unspecific' signals from S. cerevisiae array elements compared to 'specific' signals from the corresponding S. pombe array elements. Array elements for the following S. cerevisiae genes were used (increasing similarity; corresponding S. pombe genes in parentheses): HDA1 (SPAC8C9.06c); RPL18A (rpl18-1); CDC2 (cdc6); CDC19 (pyk1); RPL18A (rpl18-2); URA7 (SPAC10F6.03c); RPL27A (rpl27-2); HOG1 (sty1); YPT1 (ypt2); ACT1 (act1); HTA1 (hta1); HTB1 (htb1); ACT1 (act1).
Figure 7
Figure 7
Reproducibility of signal ratios and intensities. (A) Scatter plot showing the reproducibility between two biological repeats of an experiment where cells grown in minimal or rich media were directly compared to each other. The plot represents 4245 genes that gave measurable data in both experiments. The CV for the repeated experiment shown here is 5%. Just one gene shows an ~2-fold difference in ratios between the two experiments (just outside the outer lines). (B) Autocorrelation plot showing the distribution of Cy5 and Cy3 signal intensities from a single self-self experiment. Median signal intensities minus median local background intensities were determined and signals from replicate spots were averaged. Grey spots: data from 790 spots that were flagged 'absent' during analysis or initial data processing (see Methods). Blue spots: data from 4260 spots that were retained for evaluation. All the signal intensities from the blue spots are <2-fold different from each other (within outer lines).
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Brown PO, Botstein D. Exploring the new world of the genome with DNA microarrays. Nat Genet. 1999;Suppl 21:33–37. doi: 10.1038/4462. - DOI - PubMed
    1. Lockhart D, Winzeler E. Genomics, gene expression and DNA arrays. Nature. 2000;405:827–836. doi: 10.1038/35015701. - DOI - PubMed
    1. Young R. Biomedical discovery with DNA arrays. Cell. 2000;102:9–16. - PubMed
    1. Schena M, Shalon D, Davis RW, Brown PO. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science. 1995;270:467–470. - PubMed
    1. Shalon D, Smith SJ, Brown PO. A DNA microarray system for analyzing complex DNA samples using two-color fluorescent probe hybridization. Genome Res. 1996;6:639–645. - PubMed

Publication types

MeSH terms

LinkOut - more resources