Team:USTC Software/Database

Basic Description

The Database part is designated to build up the bridge combining the abstract parameters yielded by the System Identification process and the concrete needs from users via SBML. Therefore, corresponding to its transitive role here, the inputs and outputs of this part is similarity characterized.

The Functions

1.Filter all matching reactions from the database via matching suitable parameters of all terms of the ODE array, provided by preceding GA and PSO part;
2.Yield a complete list of substances contained in these reactions;
3.Sort reactions by reaction type, reactant/product’s type or number;
4.Make it possible for users to expand the database in directed form.

Abstract and approximation

1.All bio-chemical reactions on the genetic level are characterized in a finite set of reaction types, specifically as listed in Media:keneticlaw.png here;
2.Each term of the ODE can be mapped to part of a prototype of reactions (including all info from reactants and partial info from products);
3.Space-independent and Compartment-independent.

Data Structure and Organization

struct reactor{
	int rnum;//number of reactants
	char **ract;//names of reactants
};

struct product{
	int pnum;//number of products
	char **prdt;//names of products
};

struct reaction{
	int num;//reaction number
	struct reactor RACT;//info of reactants
	struct product PRDT;//info of products
	int type;//reaction type as listed in keneticlaw.pdf
	double *para;//all parameters
	char note[MAXLEN];//self-defined notes
};

Visualization

Instructions

{ 1.Input the tolerance of deviation of stored reactions from idea identified parameters in percentage (temporarily confined to 0% to 1000%);
2.How many results counting down from the top of the list do you want to see? Input an integer no bigger than 10;
3.Or you may tell the substances’ name you want to include just in string form less than 50 bytes.

Independent test in console application

Outputs

dbt0.dat, dbt1.dat …, dbt9.dat: all reactions contained in best ten models
sub0.dat, sub1.dat …, sub9.dat: all related substances respectively related in the ten sets of reaction

Defects

1.Not ready for repeated searches at one traverse of the original database;
2.Better performs if parameters could be interchanged though RAM/ROM;
3.Limited forms of reaction types;
4.If ordered itself, the database will be capable to offer higher searching efficiency;