Protein Lysate Microarrays

1
Protein Lysate Microarrays

• Clay Scott
• Ryan McConnell
• Shannon Neeley

2
Biological and Technical Background
3
Motivation

• DNA/RNA microarrays are used to determine gene
  expression
• Proteomic profiling can help yield more direct
  answers to biological questions
• Those molecules that can answer our questions are
  proteins, not mRNA
• Biological effector molecules
• Diagnostic markers
• Pharmaceutical targets

4
History of Proteomic Profiling

• 2D-PAGE (two-dimensional polyacrylamide gel
  electrophoresis)
• Introduced in 1975
• Semiquantitation of most abundant 1000 spots
• Problems with identifying spots with a particular
  protein
• Microarray formats for proteomic profiling
• More recent development (papers published in
  2001)
• Robotic spotting of antibodies that capture the
  protein molecule to be assessed

5
Reverse-phase protein lysate microarrays

• Opposite configuration from previous microarray
• Samples assessed are robotically spotted
• An antibody is used to measure amount of a
  particular protein present in the sample
• Limitation Measure one protein per slide
• Advantage All samples analyzed side-by-side in a
  single array
• Can compare protein levels across samples rather
  than samples across protein type

6
Protein lysate preparation

• Lysis The dissolution or destruction of cells
  by the action of a specific lysin that disrupts
  the cell membrane
• Lysate The cellular debris and fluid produced by
  lysis

7
Protein Lysate Array Design

• Each microarray is a glass slide containing
  particular protein samples from different
  patients
• 18 spots for every sample
• 3 replicates
• 6 dilution levels (with dilution factor of 2)

8
Western Blottingused to screen specificity of
antibodies

• Choose Antibody that will bind to protein

9
Detection of protein on microarray

• The slide is exposed to the antibody
• Antibody binds to the protein, depending on how
  much protein is present
• Microarray is scanned to form an image with
  darker spots reflecting higher levels of protein
• Use two antibody detection system

10
Statistical Applications
11
From Image to Number

• Software draws a circle around each spot
• Darker spots have greater quantities of protein
• sVOL reflects the total amount of protein

Two-fold serial dilutions
sVOL (pr2)(average intensity inside
circle)-(average background intensity)
12
The Data Set

• 80 x 6 x 3 matrix of sVOL values for the caspase
  protein
• 80 patients on one microarray chip
• 6 two-fold serial dilutions for each patient
• 3 replicates for each patient
• Control serum data set for calibration and
  error-assessment purposes

Dilutions
P A T I E N T S
1
6
80
13
Potential Tasks to Undertake

• From the 18 sVOL values for each sample, extract
  a single robust number that is representative of
  that sample
• for example, the mean is not robust
• Provide an error estimate of that number

14
Saturation DetectionDevise procedure for
extracting linear portion of log(sVOL)
For highly concentrated samples, circles drawn by
computer may be too small (sVOL smaller than
expected) For large dilutions, sVOL is dominated
by background noise (sVOL tails off)
15
Fine Tuning the Segmentation Program

• Outlying sVOL values may occur even when
  saturation is not a problem
• Detecting outliers may help improve the
  segmentation software