CSC2427 Algorithms in Molecular Biology Wednesday March 1 2006 Microarrays Lecturer Timothy Hughes Scribe Desiree Tillo Outline Microarray experiments Normalization Different types of microarrays Other applications besides expression profiling Clustering and interpretation Suggested reading Traditional approaches to analyze gene expression have been a time consuming process with different experiments required for each individual gene With the use of microarray technology it is now possible to analyze thousands of genes at once The main principle underlying this approach is that there is a single identical chemistry for all 4 base pairs of DNA A microarray works by exploiting the ability of complementary strands of DNA or RNA to bind specifically or hybridize to each other under appropriate conditions Microarray Experiments One type of microarray the cDNA microarray contains DNA from each open reading frame ORF from the organism of interest spotted on glass microscope slides These probes are used to detect cDNA complementary DNA which is DNA synthesized from a mature fully spliced mRNA transcript A robot spots DNA probes onto a 25 x 75 mm glass slides coated with poly lysine poly lysine creates a positively charged matrix to allow the negatively charged DNA to adhere to the slide Each of these spots are about 100 150 microns in diameter and are 200 300 microns apart allowing approximately 20 000 spots to be easily placed on one slide Once the array is made two samples of labeled cDNA experiment and control samples see Figure 1 are pooled together and are hybridized to the probes on the array Some of the common ways to label a nucleic acid include See Figure 2 1 Random priming of double stranded DNA Primers are short segments of artificially synthesized single stranded DNA ssDNA oligonucleotides which are complementary to a segment of a DNA molecule A primer is necessary in order to create a double stranded DNA from a single stranded template In order to create a labeled strand of DNA from double stranded DNA dsDNA there are three main steps 1 Figure 1 A cDNA microarray experiment Each spot on the microarray contains DNA corresponding to an open reading frame of the organism of interest Labeled cDNAs green control sample red experiment are pooled and hybridized to the array The fluorescent patterns on the array allow the determination of differences in gene expression a The double stranded template is heated to 94 96o C in order to break the hydrogen bonds connecting the two strands of dsDNA thereby separating the strands and creating ssDNA b The temperature is lowered allowing the the primer can attach or anneal itself to the newly separated DNA strands c DNA polymerase present in the reaction mixture then copies the DNA strands Starting at the annealed primer it works its way along the DNA strand incorporating nucleotides that are complementary to the template strand These nucleotides are labeled either with a radioisotope or a fluorescent molecule fluor The end result is a new complementary strand of labeled DNA In cases when there is a complex mixture of different DNA templates instead of designing primers specific to each template random octamers are used These are short primers made up of every combination of 8 nucleotides 48 65 536 different combinations Each octamer can bind anneal to any section of DNA and a labeled cDNA pool can be made with minimal effort 2 Poly T primed cDNA synthesis Messenger RNA mRNA transcripts have a stretch of adenines on the 3 end of the transcript poly A tail A labeled complementary strand of DNA can be created by using a poly T primer complementary to the poly A tail and labeled nucleotides 3 Direct labeling Here a combination of enzymatic and chemical steps are performed to modify and add the label directly to specific nucleotides present in the target strand 2 Figure 2 Common methods for labeling nucleic acids see text for details 4 Amplification Certain microarrays require a lot of cDNA so that fluorescence can be detected after hybridization The addition of the T7 promoter to the polyT primer allows approximately 100 copies of cDNA to be generated from a single mRNA template using the polymerase from the T7 bacteriophage a virus that infects bacteria Common labels for nucleic acids for microarray experiments are the fluorescent dyes Cy3 and Cy5 A green laser 523nm excites Cy3 and emitted energy is detected by an emission filter that detects emitted energy at 557 592nm Red lasers 635nm excite Cy5 and emission is detected by a filter that detects at 650 690nm The array is scanned twice once per channel used The primary data consists of two grayscale TIFF files The image is then processed and normalized Processing includes lining up the spots on the grid flagging bad spots subtracting background and quantitating the fluorescence of each spot on the array Raw data from the scanner can be displayed in an spreadsheet and contains information about each spot including the identity of the sequence its location on the array and most importantly the intensity and the normalized log ratio which is utilized to determine the relative expression of the genes represented on the array Normalization Figure 3 shows the data from a single microarray experiment plotted in a scatter plot In the scatter plot in the left panel each point corresponds to a single spot on the array and this plot shows the differences in expression of each gene the ratio of drug and no drug or red red and green log10 green and by how much given by the log10 f luorescence intensity 3 Figure 3 Microarray data from two experiments The scatter plot on the right shows the differences in expression of identical cell cultures in this case yeast Since the cultures being compared are identical there are no considerable differences in expression between them The differences in intensity in this case are due to biological noise which is due to different genes being expressed at different levels which normally occurs in the cell Sometimes wild type vs wild type comparisons take a curved shape similar to a banana or jumping whale which may be due to effects that arise from variation in the microarray technology or the experimental methods used These can include unequal amounts of starting RNA differences in labeling efficiencies of the fluorescent dyes used and systematic biases in the measured expression levels It is for these reasons that it is important to normalize the data in order to