Protein Crystallization Database: Database Design

1
Protein Crystallization Database Database
Design

• Douglas S. Kerr Vadim Cherezov
• July 15, 2002

2
Protein Crystallization

• Screening multidimensional space
• Variables
• - Temperature
• - Precipitants (100)
• In practice, combinations of no more than 5
  variables
• limits on concentrations

3
Crystallization Database Structure

• Inventories
• Records of crystallization setups
• Records of observations

4
Crystallization Database features

• Keeps track of everything
• Flexible (not limited to only one method)
• Scoring scheme for analysis
• Power to predict

5
The basic idea

• A plate with multiple cells is used.
• One protein mixture is created for the plate
  containing
• Protein(s)
• Lipid(s)
• Buffer(s)
• Detergent(s)
• Precipitant(s)

6
The basic idea, continued

• A different screen is used in each cell.
• Observations are made over time.
• At each time, each cell is given a score
  referring to the quality of any possible crystals
  formed.

7
Stocks of the Various Materials

• Stocks may be Purchased or Created in the Lab.
• Source of Purchased Stock is recorded.
• Experimentor of Created Stock is recorded.
• The composition of each stock is recorded.

8
Stock Types

• Proteins
• Lipids (Natural Synthetic)
• Buffers
• Detergents
• Precipitants

• Chemicals (Used to create)
• Buffers
• Precipitants
• Protein Mixtures
• Screens

9
The Database Design8 Subschemata

• 7 Stocks
• Chemicals and Precipitants
• Detergents
• Buffers
• Lipids
• Proteins
• Screens
• Protein Mixtures
• Observations

10
Chemical and Precipitant Stocks

• Chemical Stocks are Purchased
• Precipitant Stocks
• Created from 1 or more Chemical Stocks.
• Purchased.
• The Source of any Purchased Stocked is recorded.
• The Experimentor creating any Created Stocked is
  recorded.

11
Chemical and Precipitant Stocks
For postscript and PDF versions of this and the
other subschemata, see http//www.cis.ohio-state.e
du/doug/Lipidat/Crystallization/.
Both Created and Purchased
Chemical Stocks
Dry (T/F)
H20/D2O
Chemical Names
12
One-to-Many RelationshipChemicals to
Chemical_Stocks
13
Buffer Stocks

• 1 Main Reagent
• 0 (?) or 1 Titrant
• 0 or more Detergents
• 0 or more Chemical Additives

14
Buffer Stocks
Many-to-Many Relationship The same Detergent
Stock may be used in several Buffers. A buffer
Stock may contain several detergents.
Both the Main Reagent and Titrant are Chemicals.
15
Many-to-Many RelationshipChemicals to
Buffer_Stocks_purchased
1
?
1
?
?
1
16
Detergent Stocks

• Purchased
• Created
• 1 or more Purchased Detergent Stocks

17
Detergent Stocks
May combine several Purchased Detergents to
create a Detergent Stock
18
Lipid Stocks

• 1 Synthetic Lipid
• Or 1 Natural Lipid

19
Lipid Stocks
Synthetic Lipid
Natural Lipid
20
Protein Stocks

• 1 Protein
• Identified by its
• Source Organism
• Organism Type and
• Extract Type
• 0 or more Detergents
• 1 or more Lipids
• 0 or more Additives
• 0 or 1 Buffer

21
Protein Stocks
Created
The Protein
Purchased
22
Protein Mixtures

• Composition
• 1 Protein Stock
• May already contain Lipid(s)
• 1 or more Lipids
• 1 or more Detergents
• 1 Buffer, in addition to one in the Protein Stock
• Crystallization Method(s)

23
Protein Mixtures
24
Screen Stocks

• 1 or more Detergents
• 1 Buffer
• 1 or more Precipitants

25
Screen Stocks
Created
Purchased
26
Observations

• Collected by Plate.
• 1 Protein Mixture
• 1 Scoring Scheme
• Different Schemes may be used with different
  plates.
• Several Wells on a Plate.
• Different Screen and/or Screen Concentration for
  each Well.
• Observations repeated over time, giving a Score
  each time.
• Different Scoring Schemes may exist.

27
Observations
Each Plate has many Wells. Each Well has many
Observations.
1 Experimentor sets up the Plate. Different
Experimentors may observe the plate.