420 likes | 661 Views
DNA Microarray. Microarray Printing. 96-well-plate (PCR Products). 384-well print-plate. Microarray. Differential Expression. Each cell contains a complete copy of the organism’s genome Cells are of many different types and state e.g. blood, nerve, skin cells, etc
E N D
Microarray Printing 96-well-plate (PCR Products) 384-well print-plate Microarray
Differential Expression • Each cell contains a complete copy of the organism’s genome • Cells are of many different types and state e.g. blood, nerve, skin cells, etc • What makes the cells different ? • Differential gene expression, i.e., when, where and in what quantity each gene is expressed • On average, 40% of our genes are expressed at any given time
Functional genomics • The various genome projects have yielded the complete DNA sequences of many organisms. e.g. human, mouse, yeast, fruitfly, etc. • Human: 3 billion base-pairs, 30-40 thousand genes. • Challenge: go from sequence to function, i.e., define the role of each gene and understand how the genome functions as a whole.
Central Dogma • The expression of the genetic information stored in the DNA Molecule occurs in two stages: --transcription, during which DNA is transcribed into mRNA; --translation, during which mRNA is translated to produce a protein. • DNA mRNA Protein cDNA Arrays Tissue Arrays
Microarray Gene Expression Image
A Better Look
Cy5 Cy3 Image Analysis & Data Visualization Cy5 Cy3 log2 Cy3 Cy5 Experiments 8 4 2 fold 2 4 8 Underexpressed Overexpressed Genes
New Data ScanAlyze/GenePix Cluster Database Data Selection SOM K-means SVD Complete Data Table (cdt) SpotList
Ovarian Tumor Study M. Schaner Samples that should Cluster together do not
Different amounts of starting material. Pool of Cell Lines Tumor
Such biases have consequences: • Plotting the frequency of un-normalized intensities reveals the differential effect between the two c hannels.
How do we deal with this? Normalization: In general, an assumption is made that the average gene does not change. You must understand your experiment and data to judge whether that assumption is a good one. Usually true for gene expression experiments, but not necessarily for aCGH or chromatin IP. Generally true for large arrays, but not for small " boutique" arrays.
Normalization : The R-I Plot • Data may have an intensity-dependent structure. • Plot log2(R/G) vs. log10(R*G) to reveal this • Reveals : • variance in log ratios is greater at lower intensities. • distribution may not be centered around zero.
Normalization: Loess R-I Plot Following Loess R-I Plot, Raw Data log2(R/G) log10(R*G)
Cluster Analysis • Cell Cycle example( Spellman 1988)
Overview of the Cell Cycle • Purpose: • To create two new cells by dividing one original cell
Cell Cycle: Key Concepts • All parts of original cell must be replicated and split between new cells • Each step must occur in precise manner and timing for successful cycle, and is strictly regulated • mRNA and proteins for cell cycle genes are found at varying levels at different points of the cycle • Mutations causing malfunction in regulation can result in cancer
Cell Cycle: Basic Description http://www.bmb.psu.edu/courses/biotc489/notes/cycle.jpg