Proteogenomics

1
Proteogenomics
2
The Problem
3
Diseased organs and tissues produce tens of
thousands of peptidesMost of these are released
into the blood and appear in urineOnly a few
hundred are measured for diagnostic purposes today
4
Diseased blood cells or cells that appear in
blood express tens of thousands of mRNAs, some
with diagnostic value mRNA expression correlates
somewhat with protein expression mRNA expression
is easily measured with standard technology
5
The Process
6
Blood sample
cells
plasma
mRNA extraction
sample preparation
quality control chip (Agilent)
quality control (gel, protein)
Affymetrix chip
mass spectrometry
7
mRNA extraction
sample preparation
quality control chip (Agilent)
quality control (gel, protein)
Affymetrix chip
mass spectrometry
Bioinformatics expression level, quality peak
detection, quality sample comparison statistica
l tests Health Eval. Sciences
8
Affymetrix chip
mass spectrometry
Mellon Biomarker core
Biomolecular core
Bioinformatics expression level, quality peak
detection, quality sample comparison statistica
l tests Health Eval. Sciences
RT-PCR real-time RT-PCR RNAse protection Northern
Western flow cytometry ELISA Luminex
9
The People
10
Biomolecular core Jay Fox Alyson Prorock
Mellon Biomarker core Mark Ross David Smalley
Health Eval. Sciences Bill Knaus Jae Lee
11
The Technology
12
Genomics
13
Monocyte Macrophage Foam Cell
14
5 hrs MCP-1
5 hrs GRO-a
PBMCs
monocytes
macrophages
macrophages
5 days M-CSF or PF4
2 days media
Blood draw
2 days LDL
2 days oxLDL
2 days mmLDL
foam cells
15
log2(absolute expression)
16
normalized across rows
17
ordinal scale
18
(No Transcript)
19
8x6 Self-organizing map of all genes
20
5x5 Self-organizing map of 142 genes
21

• Technical Conclusions
• Heat maps are good to find patterns and
  relationships
• Ordinal scale is best suited to present the
  information
• Thresholding is necessary to avoid genes with low
  expression levels (
• Self-organizing maps are informative for many
  genes, but not for a few genes
• Expression patterns of individual known genes can
  be used to fish for genes with similar expression
  patterns

22

• Biological Conclusions
• lots of genes are upregulated in monocytes by
  isolation from blood
• many genes are downregulated when monocytes
  differentiate to macrophages
• chemokines produce similar and unique patterns of
  gene expression
• native LDL and oxidized LDL have similar
  effects, which are different from those induced
  by mmLDL

23
Proteomics
24
Plasma Proteomics

• Researchers have detected 289 proteins in the
  plasma by Mass Spectrometry (MS) as of Oct. 2002
  (Anderson and Anderson, Mol Cell Proteomics,
  1845-867 (2002)).
• Problem MS is sensitive but has a poor dynamic
  range.

25
Anderson and Anderson, Mol Cell Proteomics,
1845-867 (2002)
26
FIG. 5. The declining rate of introduction of new
protein tests. The data of Fig. 4 are plotted to
indicate the rate of introduction of new protein
analytes in FDA-approved clinical tests. Anderson
and Anderson, Mol Cell Proteomics, 1845-867
(2002)
27
Solutions to the low dynamic range problem

• Remove some of the proteins present in the
  highest quantities.
• Prior separation of proteins (e.g. 2D gels,
  chromatography, etc.)
• Improve MS capabilities (MS/MS)

28
What are MP?
29
Results
Figure 1. Repeated Ultracentrifugation of
Microparticles Results in the Reduction in the
Plasma Protein Contaminants. Pellets from
Ultacentrifugation of plasma were separated by
7.5 polyacrylamide gel electrophoresis and
detected by silver stain. Lane 1, Molecular
Weight standards (10 ?L) Lane 2, Plasma Lane 3,
Pellet from 1 ultracentrifugation of plasma Lane
4, Pellet from ultracentrifugation of plasma
following single wash Lane 5, Pellet from
ultracentrifugation of plasma following two
washes lane 6, As in lane 4 except five times
the amount of protein loaded in well lane 7, As
in lane 5 except five times the amount of protein
loaded in well lane 8, As in lane 4 except
twenty five times the amount of protein loaded in
well lane 9, As in lane 5 except twenty five
times the amount of protein loaded in well, lane
10, MW marker (1 ?L).
1 2 3 4 5 6 7 8 9
10
30
15 uL of microparticle sample
31
Results

• Roughly 1000 different hits.
• vWF
• Lipoproteins
• Complement Related Proteins
• Immunoglobulin related proteins
• Other plasma proteins
• Hypothetical/Unknown Proteins
• Keratin
• Cytoplasmic
• Interesting
• Others
• Albumin minor contaminant
• Few Transmembrane proteins