CRNRepair is a framework for Automated Repair of Chemical Reaction Networks. This page contains supplementatry data from our paper:

I. Mesecan, M.C. Gerten, J.I. Lathrop, M.B. Cohen, T. H. Caldas, CRNRepair: Automated Program Repair of Chemical Reaction Networks, International Workshop on Genetic Improvement, May 2021.

bibtex reference:
@inproceedings{Mesecan:GI21,
author= {Ibrahim Mesecan and Michael C. Gerten and James I. Lathrop and Myra B. Cohen and Tomas Haddad Caldas},
title= {CRNRepair: Automated Program Repair of Chemical Reaction Networks},
booktitle={International Workshop on Genetic Improvement},
year={2021},
pages={1-8}
}

The framewok uses the Systems Biology Markup Language SBML format which is based on XML. SBML is used for describing Chemical and Biological processes, e.g. Similar to XML, SBML is composed of tags and attributes. Every tag can have 0 or more attributes and 0 or more subtags.

There are two subjects in this repository: Hailstone and Subtraction. For each set, subject 0 (H0 or S0) is the correct CRN. And, subjects from 1 through 10 are the mutants that we are trying to repair.

SBML model tag contains listOfUnitDefinitions, listOfCompartments, listOfSpecies, (possible list of species in the pool) and listOfReactions. Then, reactions are given as a subtag in the listOfReactions. Every Reaction must contain at least one of (listOfReactants and listOfProducts) and the kineticLaw.

Log files contain 3 sections:

1. The AST tree enumeration: This section is at the top of log files. The lines start with a number, followed by a colon, space and a tagname. For example, the following:
   7: species
   is the 7th node in the AST (the first species in the pool of listOfSpecies) and it refers to the first species, X1, given in the input sbml file.

2. Text format of the input CRN: After AST tree enumaration, the text format of the input CRN is given, e.g.
   List of given reactions in plain text:
reaction_1: X1 -> Y
mutant-addreactant: X2 Y X1 -> null
   This is the human readable equivalent of the input sbml file.

3. Result information for each epoch: This section e.g. starts with dashed line containing "RESULT" word in the center and contains result information for each epoch. For example, the following is the result of the second epoch used in RQ1 Subtraction test subject 6 :
   Epoch: 1 Fit.Eval: 70 Time elapsed: 1786.6 BestFitness: 0.0
BestPatch: StmtInsertion({'target': ('Input/Subtraction/CRN/S6.sbml', 24), 'ingredient': ('Input/Subtraction/CRN/S6.sbml', 11)})
StmtDeletion({'target': ('Input/Subtraction/CRN/S6.sbml', 28)})

There are 4 items reported for each epoch:

• Fitness evaluation: The number of fitness evaluations in this epoch: 70. And, on 70th evaluation the target fitness is reached.
• Time elapsed: Time elapsed in seconds: 1786 seconds.
• Best fitness: Best fitness found in this epoch: 0.
• Best Patch: Best patch may contain 1 or more mutations. In the above example, there are two mutations there:
  1. Insert a copy of AST node 11 (the first reaction in the sbml file) after the AST node 24 (reaction 2)
  2. Delete AST node 28 (the last species from the listOfReactants in reaction 2, X1).

If any epoch is successful (if the target fitness is found), the corresponding patched sbml and text files are reported.

Acknowledgments: This work is supported in part by NSF Grant CCF-1909688, CCF-1901543, and FET-1900716. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.