Homework 3
Due: Thursday, Dec 6, 8:30pm
Turn in everything electronically. However, please do not send me too many extra stuff
that I did not ask for. Read carefully below for what I am expecting.
Purpose
The purpose of this exercise is for you to get familar with microarray data and some available tools for
analyzing them.
Data and tools needed
(1) Download gene expression data cellcycle.txt from the course website (http://www.cs.utsa.edu/
~jruan/teaching/cs5263_fall_2007/hw3_files/cellcycle.txt).
The original data were obtained by two different research groups using microarray to study gene
expression at different stages of cell cycles in the Saccharomyces cerevisiae (the more common name:
baker’s yeast). (If you have forgotten what is a cell cycle, just keep reading. You don’t need to know
it to finish the homework). One study was conducted using Affymetrix microarray by Cho et. al.
(Molecular Cell, July 1998), and included 17 time points covering approximately two cell cycles. The
other study was conducted by Spellman et. al. (Molecular Biology of the Cell, December 1998) using
cDNA microarray, and consisted of three subsets of experiments, each with 10-20 time points covering
about two cell cycles. The combined data were downloaded from http://genome-www.stanford.edu/
cellcycle/.
I have processed the data with the following procedure:
i. Removed genes with more than ten missing values.
ii. Missing values for the remaining genes were replaced with uniformly distributed random numbers
between -1 and 1.
iii. Genes were ranked according to the standard deviation of their expression vectors. The top 3000
genes with the highest variability were selected. The rest were discarded.
The processed data are stored as a 3000 x 73 matrix, where each row corresponds to a gene, and each
column corresponds to an experiment (a time point). Gene names and experiment names are shown
as row and column headers, respectively. Data are provided as log ratios between gene expression
levels measured from yeast cells collected at a particular time point (if you want to know, the cells
were “synchronized” so that they were in the same “cell-cycle phases”.) and those measured from a
mixture of yeast cells collected at different time points (i.e., the sample consisted of cells from different
cell-cycle phases).
(2) Download the MeV 4.0 software for micorarray gene expression clustering from http://www.tm4.org/
mev.html (or some other clustering tools if you like).
(3) FuncAssociate web interface for gene ontology analysis (http://llama.med.harvard.edu/cgi/func/
funcassociate). You can use other similar tools if you prefer.
(4) AlignACE or MEME for motif finding. Both tools have standalone versions and web interfaces. Alig-
nACE is available at http://atlas.med.harvard.edu/. MEME can be accessed and downloaded at
http://meme.sdsc.edu/meme/intro.html. I believe both tools have some limitations on input sizes
if you use their web interfaces. So if the number of sequences you provide exceeds the limit, you may
have to download and install a standalone version on your local computer.
1
