Frequency

Figure 3.18: Fourier analysis of spatial series on spotted microarray. The same data illustrated in figure 3.17 is subjected to a Fourier analysis to extract systematic periodic signals. As shown, the frequency with the largest power is 4. That is, there appears to be a periodicity of 4 per microarray based on the position of the probe on the microarray. The magnitude of this systematic variation dominates fold changes in gene expression of the magnitude reported elsewhere as noteworthy. Consider the following image of the Incyte GEM microarray at the time this analysis was done (figure 3.19). The pattern seems to suggest that the periodicity in the reported data originated from four separate but similarly behaved sources. It turns out in this example that during the array manufacturing process, four separate pins had indeed been used to spot each microarray quadrant. It should not be surprising, then, that small but systematic physical and chemical variations in the amount of probe cDNAs that were picked up and deposited by each of the four pins are reflected and recorded in the overall expression pattern one sees here. Other possible sources of the lower-level systematic variations seen in figure 3.18 may include the frequency with which the pin assembly units are replaced during the microarray spotting procedure.

Figure 3.19: Incyte GEM Microarray circa 1999. In this photograph of a GEM microarray, note that there are four quadrants which are each spotted with separate pins.

Again, although these noise examples were taken from Incyte gem microarrays and are of the intra-chip variety, it demonstrates that data from any microarray assay should be subject to at least some scrutiny or a quality control step as we have done here to screen out inherent and systematic biases in the measurement system. Numerous manifestations and sources of data acquisition error can be immediately caught by a simple visual inspection of the scanned images of a microarray. The images of the hybridizations illustrated for three different kinds of microarrays in figure 3.20 should be highly motivating in this regard.

Figure 3.19: Incyte GEM Microarray circa 1999. In this photograph of a GEM microarray, note that there are four quadrants which are each spotted with separate pins.

Again, although these noise examples were taken from Incyte gem microarrays and are of the intra-chip variety, it demonstrates that data from any microarray assay should be subject to at least some scrutiny or a quality control step as we have done here to screen out inherent and systematic biases in the measurement system. Numerous manifestations and sources of data acquisition error can be immediately caught by a simple visual inspection of the scanned images of a microarray. The images of the hybridizations illustrated for three different kinds of microarrays in figure 3.20 should be highly motivating in this regard.

Figure 3.20: Problems in images acquired from microarrays. On the left: a contaminated D array from the Murine 6500 Affymetrix GeneChip set. Several particles are highlighted by arrows and are thought to be torn pieces of the chip cartridge septum, potentially resulting from repeatedly pipetting the target into the array [156]. On the right top: local changes in intensity due to contaminants and scratches. (Derived from http://www.mediacy.com/arraypro.htm.) Right bottom: high magnification of a scanned image of a spotted microarray. Note the different sizes and shapes of the spots [199]. 3.2.7 Biological variation as noise: The Human Genome Project and irreproducibility of expression measurements

Figure 3.20: Problems in images acquired from microarrays. On the left: a contaminated D array from the Murine 6500 Affymetrix GeneChip set. Several particles are highlighted by arrows and are thought to be torn pieces of the chip cartridge septum, potentially resulting from repeatedly pipetting the target into the array [156]. On the right top: local changes in intensity due to contaminants and scratches. (Derived from http://www.mediacy.com/arraypro.htm.) Right bottom: high magnification of a scanned image of a spotted microarray. Note the different sizes and shapes of the spots [199]. 3.2.7 Biological variation as noise: The Human Genome Project and irreproducibility of expression measurements

As of 2001, preliminary results of the Human Genome Project have provided further explanations and reasons for microarray measurements having poor reproducibility in addition to the ones iterated in section 3.2.5. We now know that there exists a relatively high frequency of alternative splicing products from each gene—see figure 3.21—and are more aware of the heterogeneity of common gene polymorphisms across individuals and populations.

Figure 3.21: Estimate of the incidence of alternative splicing in genes. While most genes have only one spliced product, many have more than one, and this can affect the the gene product detection

efficacy and rate by microarrays. (Derived from Mironov et al. [132].)

Genes, which are composed of exons and introns, have the remarkably versatile property that their exons can be selectively added or deleted during transcription, giving rise to different proteins from the same gene. This form of alternative splicing appears to play a crucial role in the cellular physiology of higher organisms. For example, in the development of Drosophila melanogaster, a single gene determines the eventual sex (male or female) of the organism, depending upon how the gene is spliced [91]. In contrast to the most simple eukaryotes such as yeast, it appears that higher organisms such as humans have a much larger number of alternative splicing products per gene. The average number of distinct transcripts per gene averages 3.0 for humans as compared to 1.1 for yeast. Current oligonucleotide microarrays only target a set of subsequences for a particular individual transcript, and full-length cDNA microarrays only target a particular individual transcript. Consequently, both these technologies will show variation in signal intensity even if the same gene is being expressed at the same level but an alternatively spliced transcript is being transcribed rather than the splicing for which the microarray was engineered. Future microarrays may need to be constructed in a manner that is exon- and intron-specific, instead of gene-specific.

Apropos the heterogeneity of gene polymorphisms, there is, on average, one single nucleotide polymorphism (SNP) in every 1000 human nucleotide bases. Polymorphic variations of the same gene can be quite common, as noted by Chakravarti [43], and as a result, a cDNA or oligonucleotide probe that is engineered for a particular variant might hybridize considerably less effectively with another variant. This could be true even if the polymorphism does not change the final protein product of a gene. Although the genetic code is somewhat redundant in the sense that 64 codons code for 20 amino acids,[4] two codons that code for the same amino acid may bind significantly differently to their complementary DNA molecule and these chemical differences might be detectable by the microarray.

Was this article helpful?

0 0

Post a comment