Data from our paper ''Effects of growth feedback on adaptive gene circuits: A dynamical understanding'', by Ling-Wei Kong, Wenjia Shi, Xiao-Jun Tian, and Ying-Cheng Lai. All the files are .mat MAT-files, which are the binary MATLAB files that store workspace variables. Please refer to our [GitHub repository](https://github.com/lw-kong/Growth_Feedback_Adaptation) for the codes that we used to generate and analyze the MAT-files here. ## save_simulations_all.mat This MAT-file contains simulation results with the 425 three-node circuits. Below are the explanation of the core variables within this file. ### Q_sturcture_all The core results are in the Matlab struct 'Q_structure_all'. It contains 425 structures, each represents one of the 425 three-node circuits from 'iffl.txt' and 'nfb.txt'. For each circuit, we run M_all = 200,000 different samples with random circuit parameters. With each set of parameters, a set of different k_g from [0,0.2,0.4,0.6,0.8,1] are tested. If there's a parameter set that has adpative performance with any of these k_g, this set of parameters would be recorded in 'Q_structure_all'. There are 5 fields in 'Q_structure_all'. And with each one of the 425 sturcutres in 'Q_structure_all', the number of rows in different fildes should be identical, and equal to the number of sets of parameters that have adaptive circuit performance (the Q-value). In other words, each row represents a set of parameters. For instance, size(Q_structure_all(1).perfor, 1) = 210. This means for the first network topology that we test, there are 210 sets of parameters (out of M_all = 200,000) that has adaptive performance. The 'perfor' field of each structure shows the adaptive performance of each set of parameters at the six different k_g tested. 1 means it's adaptive, 0 means it doesn't fit our criteria for being adaptive. The 'pararand' field of each structure shows the random number used to generate the circuit parameters. The codes for generating these circuit parameters from 'pararand' can be found in 'solve_loopNet_Latin_6.m' from our [GitHub repository](https://github.com/lw-kong/Growth_Feedback_Adaptation). ### Q_set_all This (425 times 6) matrix summarizes the number of circuits that have adaptive performance with each network topologies (rows) and k_g values (columns). In other words, this is a matrix of Q-values. The six k_g values tested are [0,0.2,0.4,0.6,0.8,1]. For instance, Q_set_all(1,1) = size(Q_structure_all(1).perfor, 1) = 210. The total number of trials tested for each entry is M_all = 200,000. Thus, the number of each entry divided by 200,000 yields the success rate to randomly generate an adaptive circuit with the corresponding network topology and k_g value.