Image analysis
Gabriela Lochmanová
?• Goal of experiment
assessment
To define particular question - goal
• Hypothesis confirmation
• Investigation of a particular problem
To asses the number of replicates
• to enable statistical data evaluation
kind of detection
Workflow of sample preparation,
kind of control samples,
standards
Biological
• Involve biological diversity
- replicates represented by different samples of the
same nature
X Arabidopsis plantlets,
X the same type of tumour from different patients
Technical
• Involve an error of the method
(reproducibility)
- identical sample in several replicates
2
Biological replicates
Technical replicates
?
Experiment
• The goal of
experiment assessment • Sample
preparation
• 2 DE
• Detection
• Detection
General requirements on protein visualization
• Broad linear range of dye
intensity dependence on protein
amount in the gel
• Compatibility with following
procedures
(e.g. silver - glutaraldehyde!)
• High sensitivity
Dynamical range
= graph of dye intensity dependence (y
axis) on protein concentration (x axis)
Silver staining
– limited linearity - only upto 100 ng
of protein.
• Quantitative staining
• end-point
• lifetime
(e.g. quenching of fluorescent
dyes!)
• Detection Kind of detection
• phosphorylation: Pro-Q Diamond (pSer, pThr, pTyr), Pierce
phosphoprotein staining kit (pSer, pThr)
glycosylation: Pro-Q Emerald, Pierce glycoprotein staining
kit
• Radioactive labelling
Labelling before analysis (DIGE – CyDye, radioactive labelling)
Staining after analysis
Specific staining: post-translational modifications (PTM)
Unspecific staining: all proteins
• Visible staining: Coomassie brilliant blue (R250, G250), silver (acid x ammoniacal variant)
• Fluorescent staining: Sypro Ruby (Ex/Em = 280, 450/610 nm), Lucy (Ex/Em = 506/520 nm),
Flamingo Pink (Ex/Em = 512/535 nm), Oriole (Ex/Em = 270/604 nm), Krypton (Ex/Em = 520/580),
Deep Purple (Ex/Em = 365, 520/610 nm), Lumitein (Ex/Em = 280, 450/610 nm)
Background – 3D view
Colloidal CCB
Silver
SYPRO Ruby
• Detection
During image analysis we are
working with density of the dye.
• Detection
Protein spot – 3D view
Colloidal CCB
Silver
SYPRO Ruby
?
Experiment
• Goal of experiment
assessment • Sample
preparation • 2 DE
• Detection
• Image acquisition
• Image acquisition
Signals from biological samples are converted into
the digital data in black and white
• TIFF format, high resolution
Instrumentation for image acquisition
Instrument choice according to type of detection used
• Visible stains : densitometers
• Fluorescent stains: fluorescence scanners, cameras
Ex/Em spectrum has to correlate with Ex/Em characteristics of the instrument
PharosFX™ and PharosFX Plus Systems
Molecular Imager GS-800
Fuji FLA-3000
Typhoon 9200 Imager
Image Scanner III
S. Ruby: Ex/Em: 280, 450/610 nm
?
Experiment
• Goal of experiment
assessment • Sample
preparation • 2 DE
• Detection
• Image acquisition
• Image analysis
12
• Image analysis
Analysis using a specialized SW
• Comparison and evaluation of 2D gels
(visual evaluation of 2D gels is not possible)
•Human eye distinguishes
500 shades of grey
10 million colours
Visible light only at wave lenght of 380–760 nm
Almost ½ of human brain participites on
sight control.
13
• Image analysis Analysis using a specialized SW
Qualitative evaluation Quantitative evaluation / Statistical analysis
14
• Image analysis
• Spot quantity
= total intensity of a defined spot
(for evaluation gaussian visualization is used)
- corresponds with protein amount in a particular spot
15
• Image analysis Strategy according to defined goal
• Selection of limited number of significant spots• Selection of spots which significantly differ
based on certain design of experiment
– detection up- and down-regulated proteins – quantitative changes in the profile of certain spots
• treated sample x control
• time-dependent treatment
• Protein isoforms
• Post-translational modifications
1DE
The quality of image analysis corresponds to the quality of protein separation.
Prefractionation
1-DE
16
• Image analysis
Strategy according to defined goal
2-DE
izoforms
(change of 1aa in the sequence)
Level of PTM
-
+
17
• Analýza obrazu Evaluation using PDQuest
Image acquisition
Image editing – size, orientation
Spot identification
Normalization
Data analysis
Report
Matching
18
• Analýza obrazu Evaluation using PDQuest
• Spot detection wizard
– to select the parameters for detecting spots and background
filtration in gel scans
• Different gels – different parameters needed
Image acquisition
Image editing – size, orientation
Spot identification
Normalization
Data analysis
Report
Matching
19
noise streaking
fuzzy spots
low-abundant spots
• Image analysis Common anomalies present in 2D image
Goez et al. 2018 (modified)
saturated spot
spot overlap
20
Saturated spot Spot overlap Streaking
Fuzzy spot Low-abundant spot Noise
21
• Image analysis Evaluation using PDQuest
Spot detection and background filtration
• Scanset
= set of images originating from a single
gel (3 visualization of each gel)
Filtered GaussianRaw 2D image
Local changes in the background based on the intensity of present
protein spots (higher intense spots ~ higher background)
detection errors.
22
• Image analysis Evaluation using PDQuest
• Basic problem of image analysis: the differences in the spot position among gels
– the necessity to perform an alignment
•Matchset = set of gels which are compared
among each other within a single experiment
•
• Master gel = artificial gel; involves spots from
all compared gels
Image acquisition
Image editing – size, orientation
Spot identification
Normalization
Data analysis
Report
Matching
23
• Image analysis Evaluation using PDQuest
• Normalizace = compensation of variantion in spot size and intensity
among gels that is not due to differential protein expression
• condition for appropriate spot quantity comparison
• Variance caused by different factors:
• - pipetting errors during sample prep
• - handling errors resulting in sample loss during sample prep
• - sample loss during transfer from strip to gel
• - inconsistent staining among gels
• - inconsistent detection energy sources among gels during
• image acquisition
• - …..
• Normalization factor (according to selected method)
• Total quantity in analysis set
• Total quantity in valid spots
• Total density in gel image
• Specified value
• Mean of log ratios
• Local regresion model
Image acquisition
Image editing – size, orientation
Spot identification
Normalization
Data analysis
Report
Matching
24
• Image analysis
• Report
Image report Quantity Table report Quantity Graph report
Evaluation using PDQuest
?
Experiment
• Goal of experiment
assessment • Sample
preparation
• 2 DE
• Detection
• Image acquisition
• Image analysis
• Spot cutting
26
Manually (scalpel, spot picker)
• Visible stains
• Fluorescent stains
- transiluminator
Automation
• Spot cutter
• Spot cutting
OneTouch Plus spot picker
Exquest
UV-transilluminator
(Ex: 302, 365 nm)
Ettan Spot picker
Dark Reader
(Ex: 490 nm)
Xcise
?• Goal of experiment
assessment • Sample
preparation
• 2 DE
• Detection
• Image acquisition
• Image analysis
• Spot cutting • Digestion • Identification
• Goal of experiment
??
28
2-DE
• The best method of protein visualization in the form of protein spot which might be characterized its
abundance, localization, presence / absence.
• The most offen anomalies influencing image analysis:
vertical and horizontal streaking, fuzzy spots, saturated spots, low-abundant spots, spot ovelap, noise.
• Laborious, time consuming method, limited reproducibility, limited resolution (1 spot = 1 protein!).
Thank you for your attention.
Central European Institute of Technology
c/o Masaryk University
Žerotínovo nám. 9
601 77 Brno, Czech Republic
www.ceitec.eu | info@ceitec.cz