Procedures:

• Identification of biases in representative databases (eg. GenBank, SwissProt)
• Development of metrics for measurement of bias
• Create Datasets
  
• Real datasets (reflect real data)
• Simulated datasets (allow you to control  biases)
• Test effects of biases on real and simulated data
• Improvement of existing methods
  

What we hope to learn:

• Which kinds of biases exist?
• Which ones are important and which can we ignore?
• How do we make better datasets?
• How do we improve analytical methods?