Methods for testing biological network models

A computer system for testing or confirming a network model for a biological system,

• said network model comprising:
• a plurality of input constituents;
• and
• a plurality of output classes represented by boolean functions of values of perturbations or modifications to one or more of said input constituents,
• said computer system comprising:
• a processor;
• a memory coupled to said processor and encoding one or more programs, said one or more programs causing said processor to execute steps of:
• (a) assigning one or more cellular constituents to a particular output class having a defined change that best matches a measured change, wherein
• (i) said measured change is provided by data from one or more network testing experiments, said data comprising measured changes in the one or more cellular constituents from a first input state to a second input state,
• (ii) said defined change is provided by an influence matrix which defines changes of each output class from the first input state to the second input state according to the network model;
• and
• (b) determining an overall goodness of fit of said data to the network model, wherein the network model is tested or confirmed by said overall goodness of fit.

A computer system for choosing one or more network testing experiments for testing or confirming a network model for a biological system,

• said network model comprising:
• a plurality of input constituents;
• and
• a plurality of output classes represented by boolean functions of values of perturbations or modifications to one or more of said input constituents,
• said computer system comprising:
• a processor;
• and
• a memory coupled to said processor and encoding one or more programs,
• said one or more programs causing said processor to execute steps of:
• (a) identifying a covering set of input states, said covering set of input states comprising, for any pair of output classes, input states which distinguish said pair of output classes;
• and
• (b) choosing pairs of input states, each pair of input states comprising
• (i) a first input state in said covering set, and
• (ii) a second input state in said covering set,
• wherein, for any pair of output classes, said first input state and said second input state distinguish said pair of output classes,
• wherein said chosen network testing experiments comprise measured changes in one or more cellular constituents from the first input state to the second input state of each chosen pair of input states.

A method for testing or confirming a network model for a biological system,

• said network model comprising:
• a plurality of input constituents;
• and
• a plurality of output classes represented by boolean functions of values of perturbations or modifications to one or more of said input constituents,
• said method comprising:
• (a) measuring changes of one or more cellular constituents in one or more network testing experiments, which network testing experiments comprise perturbations or modifications of one or more of said input constituents from a first input state to a second input state;
• (b) constructing an influence matrix that defines changes for each output class from said first input state to said second input state according to said network model;
• (c) assigning said one or more cellular constituents to one or more output classes for which the defined changes from the influence matrix best match the measured changes provided by step (a) of each of said one or more cellular constituents;
• and
• (d) determining an overall goodness of fit of said measured changes to the network model.

A method of choosing one or more network testing experiments for testing or confirming a network model for a biological system,

• said network model comprising:
• a plurality of input constituents;
• and
• a plurality of output classes represented by boolean functions of values of perturbations or modifications to one or more of said input constituents,
• said method comprising:
• (a) identifying a covering set of input states, said covering set of input states comprising, for any pair of output classes, input states which distinguish said pair of output classes;
• and
• (b) choosing pairs of input states, each pair of input states comprising
• (i) a first input state in said covering set, and
• (ii) a second input state in said covering set,
• wherein, for any pair of output classes, said first input state and said second input state distinguish said pair of output classes,
• wherein said chosen network testing experiments comprise measured changes in one or more cellular constituents from the first input state to the second input state of each chosen pair of input states.