BIO/CSC295 2009F, Class 22: Microarrays (1)

Admin:
* Exam 1 returned.
* Reading for Thursday: Tuggle et al.  
  * Links were mailed to you
  * No response paper, just some prepared questions 
  * Email us your questions.
  * Read the conference paper and the INTRODUCTION to the journal
    article
* Two talks Thursday afternoon: 
  * Chris Tuggle on Bioinformatics 
  * David Sloan Wilson on Truth and Reconciliation for Group Selection
* And one talk Thursday at noon
  * Chem talk at noon: New technology for large scale analysis of proteins
* EC for attending As You Like It this weekend
  * Check the calendar for times
* EC for swim team home meet at 6pm
* Swing Friday in quad; Lessons at 8, dancing at 9
* Everyone should have received comments on their project proposals 
  from Sam, Vida, and some peers.
  * Email them to your peers
  * Vida will email you your reviewing resopnsibility if you email her
    to ask
* Reminder: You will be giving quick descriptions of your projects on Thursday.

Overview:
* About Exam 1
* Microarrays
* General Project Proposal Issues
* Time to work on Revised Project Proposals

Microarrays
* Why did we have you read this?  Because the paper you're reading on
  Thursday uses microarrays.
* Microarrays are used for studying gene expression.
  * When and where genes are expressed
* A cool outcome of the genomics era
* Vida says: "It will be one of the most important technologies for
  understanding health in the coming years."
  * Book alludes to this: Ways to direct therapies
* Background:
  * We're going to measure/collect mRNAs
  * In two different conditions
    * aerobically and non-aerobically
    * In normal and cancer cells
    * In different patients
* Techniques
  * mRNA is unstable, so we use reverse-transcriptase to make cDNA
    * We label some of the nucleotides (usually red on one set and green
      on another set, but GW uses blue and yellow)
  * Degrade the mRNA afterwards
  * We have complementary DNA on slides
    * Thousands of variants
    * Laid down by clever robots
  * Where there's a complementary match, the cDNA binds
  * Compare the light intensity for the two colors
  * Need lots of controls to make sure that you're using the right amount

* Questions:
  * Why are there so many different spots on the slides?
    * Because we want to check many different genes simultaneously.
    * Answers the question: "Which genes are differently expressed under
      these different conditions?"
  * Do we amplify?
    * Sometimes you amplify
    * Sometimes you just try to start with a larger sample
  * Do we have to keep the two treatments at the same number of copies?
    * You'd like to try to do so
  * How do they choose the oligos?
    * We don't know

* More info:
  * Affymetrix: technique measures absolute intensity
  * Another technique measures relative intensity

* Why use absolute intensity?
  * If you're varying lots of things, you get a lot of pairs that you'll
    have to measure
    * Control vs. variant 1
    * Control vs. variant 2
    * variant 1 vs variant 2
* Why use relative intensity
  * Biology is messy, helps balance out the messiness

So where's the bioinformatics?
* Lots of statistical analysis underlies relative expression levels
* How do you pick out patterns among thousands of things?

In the future, we hope to be able to use this technique to match the
right treatment to the right disease variant.

If we're measuring gene experession, does it matter that we're measuring
mRNA rather than functional proteins?
* mRNAs are not particularly stable, so measuring mRNA only gives
  you a particular time point; the total amount of protein may be different
* Not all mRNA is immediately expressed as proteins
* Not all proteins expressed from mRNA is active
* But it still tells you a lot

Questions
* What about introns and exons?
  * Once the mRNA is ready for transcription, they've been handled
* Can one mRNA code for multiple proteins?
  * Each piece of mRNA codes for a single protein, but same piece
    can be used to make multiple copies of that same protein
* Why don't we use proteins instead of mRNAs?
  * Because it's much easier to do with mRNA
* How do you determine which effects are direct and which are indirect?
  * It's really hard
  * You might argue that genes that are turned on quickly are a direct
    effect and those that are turned on later are indirect

The computational stuff is often much more custom than standardized.

Projects: Some issues
* The literature
* Initial analysis of data