NASA PAIR Program
PROTEOMICS COURSE
OUTLINE
Edward J. Carroll, Jr., Ph.D.
Lawrence M. Clevenson, Ph.D.
John Handy, B.S.
Virginia Hutchins Latham, M.S.
August 6-24
Morning (Carroll)
A. Introduction to the world of proteomics, definition, examples and methodology for determination of characteristics
B. Description of protein types and basic structure (primary-quaternary levels), glycoproteins and other modifications of basic unit
Afternoon
(Clevenson)
A. An Excel spreadsheet (First Steps with Excel, MS Word File)
B. Navigating
C. Entering data
D. Simple graphs/pie charts (Graphs and Statistics for Two Variable Data, MS Word file)
E. Accessing data on local area network
F. Filling series automatically
i. Consecutive numbers
ii. Arithmetic series
iii. Series of dates
G. Formulas
i. Copying with relative references
ii. Coping with absolute references
iii. Making a table
August 7
Morning
(Carroll)
A. Principles of protein analysis and characterization
B. Basis of separation and relationship to protein types
C. Introduction of the concept of microheterogeinity.
Afternoon
(Clevenson)
A. Plotting one-variable data, histograms
B. Statistical formulas and summaries
i. Mean and median
ii. Standard deviation
iii. Empirical rules
iv. Z-scores and normal distributions
Morning
(Carroll)
A. Methods of analysis of protein diversity and abundance
B. Detailed presentatioin of chromatography and electrophoresis. The practicum will include review of actual chromatographic and electrophoretic data. Electrophoresis will be demonstrated with apparatus, completed gels will be provided for analysis. The students will make measurements and set up Excell data files, prepare standard curves graphically.
Afternoon (Clevenson)
A. Plotting two-variable data, and best-fitting straight line
i. Linear data example
ii. Non-linear data example
iii. Formulas in Excel for best-fitting line
B. Choosing x and y
i. Review of straight lines
ii. Mathematical relations and statistical relations
C. Transforming non-linear data to linear
i. Review of logarithms
ii. Comparing log proteomic data versus nonlog
Morning
(Carroll)
A. Animal sperm and egg structure and fertilization and application of proteomics to current problems
Afternoon (Clevenson)
A. Plotting several graphs on the same view
B. Automating repetitious procedures; recording macros (Handout: Recording a Macro in Excel, MS Word file)
August 13
Morning
(Latham)
Mentor student in laboratory exercises to get data sets for statistics on standards and unknowns
Afternoon
(Clevenson)
A. Other models for proteomic data
Handouts:
A cross validation macro, MS Word file
A quadratic interpolation macro, MS Word file
i. SLIC attempt
ii. Cubic polynomials; multiple linear regression
B. Models with squares and square-roots
Morning
(Handy)
Overview Discussion - Digital Image Processing
Creation - Scanner
Monochrome
Gray Scale
Bi-level
RGB
Display on computer screens
Print to paper
Storage in computer files
TIFF format
Pixel data
format data
color data
Managing files
Image Filtering
Modify the image to obtain a new image
Detect features of an image
High pass
Low pass
Blurring
Edge detection
Image data as a 2D signal
Signal values assigned to a grid of positions
Signal processing 1D signal
Signal values assigned to a number line - time or position.
Description of the SpotFinder Program
Size of images to be processed
Type of filtering used: LOG(LOG**2)
Demonstration of spot finding.
Demonstration of basic image display features.
Afternoon (Clevenson)
A. Other models for proteomic data
i. SLIC attempt
ii. Cubic polynomials; multiple linear regression
B. Models with squares and square-roots
Morning
(Handy)
Detailed Discussion
Signal Processing - 1D signal
1 dimensional collection of measured values.
Usually 8 bits for values in the range 0 to 255.
Compute a new set of signal values from the original set of values.
Example filters
Filter types
Gaussian Filter
Laplacian Filter
Compound Laplacian of Gaussian Filter (LOG)
[1D] Hands on experiments seeing the effects of different filters on different kinds of data. Students run Matlab programs to compute filtered signals and display the inputs and outputs. Sample data are simple 1D examples of impulse, sinusodial, gaussian, etc.. to show the effects of the types of filters to be used.
Signal Processing - 2D signal
Separable filters - 2D filter as combination of 2 1-D filters.
[2D] Hands on experiments seeing the effects of the same types of filters as before on 2D image data.
Implementation - Software Development
Effect of large data volume on different implementations
Memory Management
Different kinds of computer memory
RAM - Semiconductor - FAST
DISK - Mechanical - SLOW
Virtual Memory
Software controls which memory is used
[Listing of very simple implementation of a filter - convolution.]
[Demonstration of memory usage.]
[Listing of a more memory efficient implementation.]
[Demonstration of memory usage.]
Description of the SpotFinder Program
Size of images to be processed
Type of filtering used: LOG(LOG**2)
Memory hog version
Memory efficient version
Demonstration of program operation.
Demonstration of basic image display features.
Demonstration of the intermediate steps of the filtered image data. [This will parallel the hands on examples previously covered.]
Afternoon(Clevenson)
A. Cross-validation, or predictive sample reuse
B. What model appears best
C. Local linear interpolation
D. Local quadratic interpolation
All
Day (Carroll/Clevenson)
A. Use of power point and a miniproject
All
Day (Carroll, Clevenson, Handy, Latham)
Mentor students in their projects, prepare final presentation
All day (Carroll, Clevenson, Handy, Latham)
Presentation day