Big picture

1. Yeast synchronization and sampling
2. RNA extraction of samples
3. cDNA prep
4. Hybridization
5. Scanning
6. GenePix feature extraction
7. Data normalization

Yeast synchronization and sampling

Day 1: before the experiment

• Inoculate yeast strains from stock in -80°C into SD medium (30-50ml) Don’t forget to put water in fridge to chill
• Incubate cultures overnight at 30°C, 225 rpm

Day 2: of the experiment

• For each strain, inoculate from mother culture into 125 ml fresh media (in 500 ml flask). The final OD should be ~0.35 (Spec factor 2). Incubate at 30° C for about 1 h (or OD reaches ~0.4—0.45). Always proceed in order: start from strain A → F

α-factor synchronization

• Add α-factor (4 µM final concentration) to each culture.
• Incubate cultures at 30 °C, 225 rpm for 1.5—2 h.
• Check cells under microscope to confirm sample has < 10% new buds and < 10% shmoos, and remaining cells are round.
• Check under light microscope if most of the cells (85-90%) at unbudded or shooming stage.

During the breaks

1. Prepare and label all needed tubes, flasks, media, ice, etc for the washing and release steps.
2. After 1 h check under the microscope the synchrony of one culture to make sure the pheromone is working.

Wash out α-factor

Once ~80% cells are unbudded or shmooing, wash out the α-factor. Set the incubator temperature ot 18° C.

• Dispense cultures into 250 ml tubes. Use a centrifuge machine (Phil Rea’s lab) to centrifuge at > 5000 rpm for 5 min at 4° C.
• Wash twice with cold water.

• Keep the pellets on ice.

  NOTE: this part should be done as fast as possible and on ice.

• Add 250—275 ml fresh SD medium (ideally at 10°—15° C) to each tube with pellet, and mix by shaking/vortexing. NB don’t forget to add Leucine supplement to lab strain YPS183.

• Get 25 ml for T0, mix immediately in cold EtOH, store in -80° C.
• Split 25 ml x 9 time points in small flasks and let grow at 18° C in shaker until sampling time.

Strains:____________________

Point	Actual time	Waiting time	Completed?
T0	0	0
T1	24	24
T2	48	24
T3	63	15
T4	87	24
T5	111	24
T6	135	24
T7	152	17
T8	176	24
T9	194	18
T10	218	24
T11	227	9
T12	251	24
T13	260	9
T14	284	24
T15	301	17
T16	325	24
T17	345	20

Point	Actual time	Waiting time	Point	Actual time	Waiting time
T0	0	0	T0	0	0
T2	48	48	T1	24	24
T4	87	39	T3	63	39
T6	135	48	T5	111	48
T8	176	41	T7	152	41
T10	218	42	T9	194	42
T12	251	33	T11	227	33
T14	284	33	T13	260	33
T16	325	41	T15	301	41
			T17	345	44

• Sampling: For each time point, record the culture’s OD 600.
• Mix sample with 20 ml of ice cold ethanol (-80°C) in a 50ml Falcon tube.
• Store in -80 °C until RNA extraction.
• These sampling tables show the three different methods used to sample cells after release from synchronization.

Supplements

• Overall protocol for yeast sampling and RNA extraction
• Yeast synchronization and sampling on 31 Jan 2007
• Yeast sampling schedule details list
• Yeast sampling schedule - contiguous 18 time points
• Yeast sampling details, 21 April 2007 - DFS

  RNA extraction of samples

Using the Qiagen RNeasy Mini Kit

• Centrifuge the sample for 3min at 5000 rpm
• Mix the pellet with a lysis buffer supplemented with β-mercaptoethanol under a chemical hood to get a final concentration ~5.108 cells/ml.
• Transfer (600 µl) into a 2ml screw cap tube with 600 µl of acid washed glass beads
• Break cells with a bead mill at Max power for 40s (2X, put on ice in between).
• Transfer the lysate to a new microcentrifuge tube and centrifuged to a Max speed.

• Transfer the supernatant to a new microcentrifuge tube and afterwards the Qiagen® RNeasy Mini Kit protocol is followed. RNA stored at –80° C

• Quality/Quantity checking of the RNA (Genequant, Amersham)

  • Blank: 100 µl of buffer (10 mM Tris-Cl, pH 7.5)
  • Sample: 1 µl of RNA extract + 99 µl of buffer

cDNA preparation, hybridization, and scanning

• Junhyong Kim Lab Expression Microarray Protocol
• cDNA synthesis macros (Excel)

Other microarray protocols

• The Big Microarray Book
• Peter White’s array protocol (UPenn Functional Genomics Core) (and other protocols)
• Amino Allyl Labeling and Hyb. Indiana U.
• Stanford Microarray Protocol
• Dunham protocol (Princeton)

Component resources

• Invitrogen SuperScript III amino-allyl indirect cDNA labeling
• Corning glass slide hybridization reference
• DyeSaver 2 manual
• AlexaFluor reactive dye decapacks for microarray applications

GenePix feature extraction

• GenePix WebSite: moleculardevices.com
• GenePix 6.0 Manual
• GenePix flowchart from UCSF (http://arrays.ucsf.edu/protocols/)
• Michael Eisen’s ScanAlyze manual

Procedure

In the interest of time, there is no need to perfect the mask for every slide. Try to obtain a good GPR file for each slide in about 20 minutes.

1. Load images from file menu, selecting either a single two-channel TIFF or two single-channel TIFF image files

2. Register images to align both single channel images

3. Set the ratio formulation (… button or in preferences): green/red, red/green

4. Feature Recognition for each slide

   1. Load gal file (GenePix array list) from file menu (or clear all flags for existing mask)
   2. Align array list: Drag the entire array list so that the first block lines up properly with the top left block of spots
   3. Auto feature extraction: Use “align blocks” button (F8 runs automatic alignment)
   4. Go to feature mode (F). Be primarily concerned with spot masks aligned to the wrong spot or to some noise (background fluorescence, hair on the slide, etc)
      • Mouse over and select feature(s) of interest
      • Translate and scale individual features (control key and arrows)
      • Only flag features that are BAD (A). A BAD spot is one whose mask definitely does not contain signal for the appropriate spot
   5. Save GPS (GenePix settings) file. This filename will appear at the bottom right of the application window and should match the barcode.

5. Analysis: background, generate GPR table, scatterplot, etc.

   • Choose a background subtraction method (default method is local)
   • Press the analysis button
   • Perform identical background correction for all slides that will be analyzed together

6. Save GPR (GenePix results) file

   • If everything is consistent, you will see the word “linked” at the bottom of the GenePix window, and the settings file will have the same barcode as the gpr file.

Notes

• Make sure scanner settings are the same for slides we are going to analyze
• GenePix histogram view
  • differences between channels indicates intensity dependent differences; i.e. use an intensity dependent normalization (lowess)
  • if you see issues here, go back to scanner and change setting to fix linear intensity span

Data normalization

1. Populate a map file describing the biological conditions of each scanned microarray slide.

   FILL IN DETAILS HERE

2. Call the normalizeArrays.py Python script: normalizeArrays.py --in <indir> --out <outdir> --map <mapfile>

Data Normalization Procedure Implemented in normalizeArrays.py

The following procedure is applied independently to each GPR data file provided in the mapfile (or indir if not map is provided).

Each GPR file contains a vector of raw intensity data, for each channel, whose length is the number of spots on a microarray slide (~16,000). Each data point is a 12 bit integer representing the median intensity of a spot on one channel. (Each spot contains approx. 400 pixels, each of which is an integer. The median summarizes this as one value per spot per channel.) In addition to the red and green foreground channels, there are also red and green background channels, which contain the median local background for each spot. Thus each spot is associated with four data points, two foreground, two background.

1. Log transform all four data histograms.

2. Background filter: Remove spots below background cutoff. A spot is removed if ch1 intensity is less than ch1 background cutoff and ch2 intensity is less than ch2 background cutoff. The default cutoff is the median of the background intensity distribution. This can be changed at runtime using --bgpct, or turned off completely with --nobgfilter. After this point the program only works on the two foreground channels.

3. Mean center the intensity histogram of each channel.

4. Correct for dye-intensity bias using lowess.

5. Correct for relative intensity (location) differences by subtracting half the median of log ratios from ch1 and adding half the median of log ratios to ch2.

6. Within slide normalization: scale the red channel to match green channel using linear regression.

7. Filter genes with high replicate variance. As a quality control check, we want the replicates of each unique oligo to have a comparable intensity on each channel. Thus a gene’s is discarded if the variance of its repliate normalized values on either channel exceed a threshold of 2 standard deviations of the intensity range of all spots on the current slide.

8. Average the replicate spot values for each unique oligo per channel. To maximize the recovery of gene data, the replicate spot values of each unique oligo from an individual channel are averaged. E.g., if an oligo has two replicate spots, with two data points on the red channel, these are averaged into one. If this oligo has one “NA” and one data point, the single data point is used.

9. Compute log ratios by taking difference of paired values from each channel.

10. Between slide normalization: Scale all log ratio distributions by regressing against an ideal random normal distribution, N(0,1.5).

11. Median center log ratio distribution for each slide. This removes any location shift due to non-zero intercepts estimated in the regression normalization.

12. Average technical replicate data vectors (group by (strain, stage) pairs), as well as any further arbitrary grouping specified using the —group flag (e.g. average all slides with the same cDNA processing date).