Unformatted text preview:

1Bio39/139: MicroarraysMichael L. Whitfield, Ph.DDepartment of GeneticsDartmouth Medical SchoolMolecular PhenotypingBreast Cancer; Sørlie et al. 2003 PNASLecture OutlineExperimental DesignBiological QuestionMicroarray PlatformBiological vs. Technical ReplicatesMicroarray platformsData miningCluster analysis and pattern recognitionGene Ontology annotationsMicroarray Platforms• DNA samples mechanically printed onto glass slides(either PCR products or oligonucleotides)• Ink-jet printing of DNA (Agilent Technologies)• Short oligonucleotides (~25 nts) synthesized in situby photolithography (Affymetrix)• Oligonucleotides synthesized in situ using masklessphotolithography (Nimblegen).•Oligonucleotides synthesized in situ usingmicroelectrodes (CombiMatrix)Microarray Design• First spotted microarray were produced with PCR products ofentire ORFs (yeast) or from clone inserts (Human, ResearchGenetics).– Significant cross-hybridization potential– Unable to distinguish which strand is expressed• Almost all new microarray designs are produced usingoligonucleotides– These are more specific than PCR’ed ORFs or clones– Are able to distinguish strand specificitySpotted Microarrays• PCR products or 70mer oligonucleotidesrobotically spotted by capillary action ontoglass microscope slides• Up to 44,000 features per slide• Hybridize two samples simultaneouslyCost $135 - $150/array2cDNA MicroarraysPros and ConsLow cost: $35 - $135 per microarray but also lower quality thancommercial arrays so you often need to hybridize more samplesFor spotted cDNA arrays there is substantial startup associatedwith the PCR; oligonucleotides are making this less of an issue.These arrays often require more input RNA for labeling andhybridization but new labeling methods have brought this down soone can start with small amounts of material.Agilent MicroarraysAgilent Human 1A oligo array22,000 elements~17,000 genesAgilent Microarrays• 60 mer oligonucleotides are 8-fold moresensitive than 25mers.• Inkjet technology - no masks are required socustom arrays can be produced for a relativelyinexpensive price.• Custom arrays can be produced fromsequence files, generally within a few days.Agilent MicroarraysAgilent Microarrays3Affymetrix MicroarraysIn situ synthesis• Uses photolithography to build25 nt oligos on chip using a setof masks.• Each set of masks are expensiveto make (~250K).Affymetrix MicroarraysCost $450/array•Typically hybridize only one sample per chipNimblegen Microarrays• Maskless array synthesizer using a DigitalMicromirror Device (DMD).• Up to 195,000 60mers or 390,000 24mers.• Inexpensive custom arrays.• New array design can be made every 3 hrs makingit possible to perform iterative microarray design.Nimblegen Microarrays• Uses photolithography• No mask necessarymaking it relativelyinexpensive to arrayany genome.• Inexpensive customarraysPlatform comparison0.2 - 1 µg10 - 20 µg$180 - $450Per arrayAgilent1 - 5 µg50 - 100 ngN/A$450 perarrayAffy0.5 - 2 µgtotal RNA20 - 50 µgtotal RNA$35 - $150per arraycDNAAmp.LabelingDirectLabelingCostExperimental Design4Microarray experimental designDesign gene expression experiments that allow us to survey all ofthe biological processes occurring in a cell, tissue sample, organismwithout prior knowledge or assumptions about the biological processunder study.Design the experiment to identify groups of genes that show similarpatterns of gene expression - often these genes are involved insimilar biological processes.Observation of these global gene expression patterns often timesgenerate new hypotheses that can be tested experimentally.Genes with similar patterns of gene expression often are regulated bysimilar mechanisms. Computational approaches are providinghypotheses about the cis-acting elements and trans-acting factorsregulating groups of genes that can now be tested experimentally.BiopsyPool of Cell LinesAgilent Human 1A oligo array22,000 elements~17,000 genesTwo-color experimental design1. Ratios are more robust than absolute intensities;variation from hybridization and spot size/DNAcontent are reduced.2. In the case of a two-color experiment,particularly with spotted arrays, the relativefluorescence at each spot is what is mostimportant.Experimental Design1. Experimental design dictates what you can dowith your data.2. Good normalization and processing areenabled by good experimental design.3. Your experimental design also facilitatescertain comparisons between samples andprovides the statistical power you need forassigning confidence limits to individualmeasurements.Comparing Two ClassesComparison will tell us about the differences ingene expression between a breast tumor andnormal breast tissueBreast TumorNormal Breast TissueComparing between two classesHow to compare a breast tumor and an ovariantumor?Breast TumorNormal Breast TissueOvarian Tumor Normal Ovary5Using a common referenceBreast TumorOvarian TumorNormal BreastNormal OvaryUsing a common referencesample can allow indirectcomparison between manydifferent samplesTotal arrays used: 4With dye flips: 8Advantages of a reference-based design• Easily extensible• Simple interpretation of the results• Requires less RNA per experimental sample• Less sensitive to bad RNA samplesUniversal Reference RNAControls for intra- and inter- microarrayvariability such as microarray design,spot size and DNA deposition, RNApreparation and labeling, hybridizationconditions and data acquisition.Goal of providing signal at every spot onthe microarray that is expressed in yoursample of interest.This can be pooled RNA from cell linesor in the case of a time course, anuntreated population of cells in which thetime course will be done.Dis-advantages of a reference-based design• Reference sample is measured over and overagain.• Biological samples are measured only once ortwice.• All comparisons are indirect, relying on thereference sampleExperimental Design: Two-color Fluorescent HybridizationReverse transcribe each sampleusing a different fluorescentnucleotide (Cy3 or Cy5)ExperimentsExtracting Data200 10000 50.00 5.644800 4800 1.00 0.009000 300 0.03 -4.91GenesCy3 Cy5Cy5Cy3log2Cy5Cy3       6ExperimentsGenesGene name 1Gene name 2Gene name 3 etc……..Name Process Function LocationThe Display Connects Expression Data with BiologyIn the Stanford implementationthe gene names are linked toA suitable database: SGD for yeast,SOURCE for human


View Full Document

DARTMOUTH BIOL 039 - MICROARRAYS

Download MICROARRAYS
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view MICROARRAYS and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view MICROARRAYS 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?