Code Readme
- Sensitivity Analysis
i. DriverBasic_sens.m -- sens.mat and creates ranked sensitivities and time-varying sensitivities figures (Fig. S2A-B)
ii. Covariance.m -- saves Correlations.mat
iii. SensHeatmap.m -- creates heatmap figures (Fig. S2C-D)
- Optimization
a. [Individual Optimization]{.underline}
i. mainLogistic.m -- optimizes 4 parameters for each patient, saved into PARS_4vary.mat
ii. par_boxplots.m -- creates parameter boxplots (Fig. S3A)
iii. ModelvData.m -- creates plot of model output versus data for all patients with 4 varying parameters (Fig. S3B)
b. [Nested Optimization]{.underline}
i. NestedOpt.m -- nested optimization for training set, saves PARS.mat and rhovarphi.mat
ii. mainLogistic.m -- individual optimization for testing set patients using two uniform parameters in rhovarphi.mat
iii. PlotResults.m -- plots individual fit for training (Fig. S4A-B) and testing (Fig. 1A-B) patients. Also plots stem cell proportions (Fig. S5)
ModelvData.m -- plots model output versus data for patients in training set (Fig. S4C) and testing set (Fig. 1C)
iv. par_boxplots.m -- creates boxplots for ps and alpha for training (Fig. S4D) and testing (Fig. 1D) data sets
v. psvsalpha.m -- Fits a curve through ps versus alpha for training (Fig. S4D) and testing (Fig. 1D) data sets
c. [Cycle by Cycle Optimization]{.underline}
i. CycbyCycOpt.m -- optimizes each cycle individually for each patient using two uniform parameters in rhovarphi.mat
ii. boxplot_CyctoCyc.m -- plots boxplot of ps changes from cycle to cycle for training patients (Fig. 8A)
- Forecast
i. parCPD.m, parCPD_2.m, parCPD_3.m -- creates cumulative probability distributions of changes in ps from cycle to cycle (Fig. 8B)
ii. mainLogistic.m -- fits curves through ps versus alpha for each cycle, saves into psalphafit.mat
ConfidenceInt.m -- plots ps versus alpha with fitted curve and 95% confidence interval (Fig. 8C)
iii. mainForecast.m -- forecasts response for each patient, given a cycle number and plots the results (Fig. 2)
- Simulations
a. [No Induction]{.underline}
i. NoInductionSims.m -- simulates IADT with and without induction, as well as continuous ADT. Saves time to progression (TTP) into TTP.mat and plots simulation results for each patient (Fig. 3G)
ii. SwimmerPlots.m -- creates swimmers plot comparing time on treatment for each patient (Fig. 3A-D)
iii. KapMeier.m -- plots Kaplan-Meier comparing TTP with and without induction and with continuous ADT (Fig. 3E)
iv. TTP_Scatter.m -- compares TTP between each of the simulations for all patients (Fig. 3F)
b. [Alternative Thresholds]{.underline}
i. ThresholdSims -- simulates IADT with alternative thresholds. Saves time to progression (TTP) into TTP.mat and plots simulation results for each patient (Fig. 4F)
ii. SwimmerPlots.m -- creates swimmers plot comparing time on treatment for each patient (Fig. 4A-C) for each threshold
iii. KapMeier.m -- plots Kaplan-Meier comparing Bruchovsky IADT protocol and alternative threshold IADT(Fig. 4D)
iv. TTP_Scatter.m -- compares TTP between each of the simulations for all patients (Fig. 4E)
c. [Induction Docetaxel]{.underline}
i. DOCSims.m -- simulates IADT with and without induction docetaxel for each patient (Fig. 5C-D)
ii. KapMeier.m -- plots Kaplan-Meier comparing TTP Bruchovsky IADT protocol IADT with induction docetaxel (Fig. 5A)
iii. PlotTTP.m -- compares TTP between IADT alone and IADT with induction docetaxel (Fig. 5B)
d. [First Cycle Docetaxel]{.underline}
i. FirstCycDOCSims.m -- simulates IADT with and without docetaxel after first cycle. Saves TTP into TTP.mat and plots simulations results for each patient (Fig. 6D)
ii. boxplot_BenvNoBen.m -- comparison of ps values between patients who do not benefit from docetaxel and those that do (Fig. 6A)
iii. KapMeier.m -- plots Kaplan-Meier comparing IADT with and without docetaxel after the first cycle. Stratifies based on ps value (Fig. 6B)
iv. TTP_Scatter.m -- compares TTP between each of the simulations for all patients (Fig. 6C)
- Leave-One Out
a. [Leave-One Out]{.underline}
i\. RunAll.m -- runs leave-one out study for each patient individually
b. [ROC Analysis]{.underline}
i. ROC.m -- plots ROC curves for all patients simultaneously (individual ROC results saved in ROC.mat)