ScoreExpression

Public Methods

• GenerateScores(DataSet, DataGrouping) - defaults to ignore zeroes
• GenerateScores(DataSet, DataGrouping, bool ignore_zeroes) - Generates 2 columns
  • [cat_Oexpr] - "1" if this observation's group 1 values are overexpressed as a categorical variable, "2" if it's group 2 values are overexpressed, "0" if they are tied
  • [cont_Oexpr] - "1" if group 1 values are overexpressed as a continuous variable, "2" if group 2 overexpressed, "0" if they are tied

Remarks

• Overexpression as a categorical variable is determined by taking the ratio between nonzero(or nonmissing) values to the total possible values in that group and determining which group's ratio is larger.
  • For example, if row X had 3 nonzero values out of a possible 5 in group 1, and 2 nonzero of a possible 5 in group 2, it would be overexpressed in group 1

• Overexpression as a continuous variable is determined by adding all values in a given group together, and then dividing by the total number of nonzero(or nonmissing) values in that group, and comparing these two ratios, the larger ratio is the overexpressed one.
  • For example, if row X had 5 nonzero values which sum to 100 in group 1, and had 5 nonzero values which sum to 50 in group 2, it would be overexpressed in group 1

• If "ignore zeroes" is false, zeroes are treated as being expression levels in the categorical expression