Skip to content
Branch: master
Find file History
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
..
Failed to load latest commit information.
Results_Summer_2017
Revisions_Fall2018
Trying_Out_Hyp_Results_Fall2017
AlternateMinimums.m
AlternateMinimums_Fall2018.m
AlternateMinimums_Fall2018_forMULTIC.m
CombineTrainTestData.m
Cost.m
CrossValidate.m
CrossValidate_CVSTUDY.m
FigureSettings.m
GenerateWildBootData.m
GetAvailableSamples.m
GetBestReg_BYo.m
GetCellCtsCELL.m
GetCellCtsMATRIX.m
GetCellCtsVEC.m
GetCombinations.m
GetCombinations_CVSTUDY.m
GetConfidenceIntervals.m
GetFullWellIndices.m
GetK14.m
GetNormFactors.m
GetTrainTestDW.m
GetTrainTestDW_CVSTUDY.m
GetWarmStart.m
GetWarmStart_MULTIC.m
IdentifyDynamics.m
IdentifyDynamics_Fall2018.m
IdentifyDynamics_Fall2018_forMULTIC.m
Main_PhenotypeDynamics.m
PREPSeeDynFit.m
PlotConfidenceIntervals.m
ProcessAMIter.m
ProcessCVData.m
ProcessCVData_CVSTUDY.m
README.txt
ReadDeathOption.m
Script_Revisions_Fall2018_2pMODELvsTestData.m
Script_Revisions_Fall2018_ENSEMBLEonestep.m
Script_Revisions_Fall2018_PvalsDynParam.m
Script_Revisions_Fall2018_SysIDMULTIC.m
SeeCV.m
SeeDynFit.m
SeePrePostData.m
SetupAM.m
SortBySeries.m
SortRawData.m
SplitIntoTrainTest.m
StackDynamicsParametersHorizonally.m
Test_SetUp.m
Timeseries_Raw_15wells.xlsx
Timeseries_Raw_4wells.xlsx
argmin_A_Xfix.m
argmin_A_Xfix_Fall2018.m
argmin_X_Afix.m
argmin_X_Afix_Fall2018.m
getPredBandENSEMBLE.m
getPredCellONEMODEL.m
getPredOneStepENSEMBLE.m
getPredULBandONEMODEL.m
getPredvsSampSTAT.m
plotBandENSEMBLE.m
plotDataSampENSEMBLE.m
plot_4well_test_data.m
plot_4well_testdata.fig
seeEvol_of_A_inAM.m
try_AlternateMinimums.m
try_CrossValidate.m
try_IdentifyDynamics.m
try_argmin_A_Xfix.m
try_script_hyp_imppres_fall2018rev.m
try_script_hyp_nov9.m

README.txt

Author: Margaret P. Chapman
Date: June 23, 2017
Purpose: Description of functions for phenotype dynamics code

AlternateMinimums.m : Executes alternating minimization to optimize dynamics with (lambda, mu, gamma), specific agent.

argmin_A_Xfix.m : Finds dynamics (A) that minimizes the cost function for fixed filled-in cell counts (X).

argmin_X_Afix.m : Finds cell count matrix (X) that minimizes the cost function for fixed dynamics (A).

CombineTrainTestData.m : Forms pre-optimization data for all (train + test) wells, death option o.

Cost.m: Defines cost function for alternating minimization.

CrossValidate.m : Gets best training parameters via cross validation.

FigureSettings.m : Sets figure settings for nice plots.

GenerateWildBootData.m : Resamples residuals via bootstrap technique proposed by Wu 1986.

GetAvailableSamples.m : Returns column indices of available samples.

GetBestReg_BYo.m : Extracts best (lambda, mu) for each death option, based on cross-validation.

GetCellCtsCELL.m : Gets cell counts for all samples given specific death allocation option & agent.

GetCellCtsMATRIX.m : Computes cell count matrix for cvx optimization, agent a.

GetCellCtsVEC.m : Gets cell count vector for one sample given live & dying phenotype tallies, fraction dying, g+, & g-.

GetCombinations.m : Sorts combinations of (lambda, mu, death option) into an easy-access matrix.

GetConfidenceIntervals.m : Computes upper and lower bounds of a 95% confidence interval for each dynamics parameter.

GetFullWellIndices.m : Identifies indices of wells with phenotype & death timeseries raw data fully available.

GetK14.m : Converts phenotypes to K14+, K14-.

GetNormFactors.m : Computes normalization factors for each term in cost function.

GetTrainTestDW.m : Returns train test data & wells for cross-validation, agent a.

GetWarmStart.m : Computes 'warm' starting condition for data in alternating minimization.

IdentifyDynamics.m : Performs all model identification computations, agent a.

Main_PhenotypeDynamics.m : Main script for phenotype dynamical model.

PlotConfidenceIntervals.m : Plots 95% confidence intervals for all dynamics parameters and all agents.

PREPSeeDynFit.m : Sorts test sample & prediction into data structures for later plotting.

ProcessCVData.m : Precomputes train & test matrices for each death option.

ReadDeathOption.m : Returns value of gamma_hi & value of gamma_lo given sample & death option.

SeeCV.m : Shows cross-validation error vs. (lambda, mu, death option), agent a.

SeeDynFit.m : Shows how well dynamics identified on 15-well data fits to 4-well data.

SeePrePostData.m : Shows distribution of post-optimization vs pre-optimization data.

SetupAM.m : Performs one-time calculations for alternating minimization.

SortBySeries.m : Separates phenotype & death timeseries.

SortRawData.m : Sort raw data according to agent & time point.

SplitIntoTrainTest.m : Split full data matrix into train & test.

StackDynamicsParametersHorizonally.m : Arranges parameters by replacing diagonals with cell division gains and transposing each column of the dynamics matrix.

Test_SetUp.m : Test script for set-up.

Timeseries_Raw_4wells.xlsx : raw 4-well phenotype & death timeseries data for testing dynamics

Timeseries_Raw_15wells.xlsx : raw 15-well phenotype & death timeseries data for identifying dynamics

---------------------------------------------------------------------------
Revisions - Fall 2018

SYS ID WITHOUT REDUNDANT SECOND LOOP
AlternateMinimums_Fall2018.m : ensures that we break out of optimization loop if it's successful
argmin_A_Xfix_Fall2018.m : redundant second loop corrected
    -specified w/ "Fall2018" so readers would have access to the original code as well
argmin_X_Afix_Fall2018.m : redundant second loop corrected
    -specified w/ "Fall2018" so readers would have access to the original code as well
IdentifyDynamics_Fall2018.m : facilitates AlternateMinimums_Fall2018.m

CONVERGENCE OF AM ALGORITHM
seeEvol_of_A_inAM.m : shows how the dynamics parameters evolve at each iteration of AM algorithm

ENSEMBLE MODELING
Script_Revisions_Fall2018_ENSEMBLEonestep.m : uses ensemble modeling to make predictions
getPredBandENSEMBLE.m : provides a band around the median of the trajectories
getPredOneStepENSEMBLE.m : returns predicted state trajectories under all bootstrap models, agent a
getPredvsSampSTAT.m : quantifies concordance/disconcordance between two time-series data sets via two-way ANOVA
plotBandENSEMBLE.m : shades the space between an upper curve and a lower curve, and plots a middle curve
plotDataSampENSEMBLE.m : plots data samples over the time horizon

P-VALUES FOR DYNAMICS PARAMETERS
Script_Revisions_Fall2018_PvalsDynParam.m : computes a p-value for each dynamics parameter using bootstrapped samples

STARTING FROM MULTIPLE INITIAL CONDITIONS
Script_Revisions_Fall2018_SysIDMULTIC.m : does sys id from multiple initial conditions
GetWarmStart_MULTIC.m : adds noise to inital cell counts matrix
AlternateMinimums_Fall2018_forMULTI.m : does alternating mimization for multiple initial conditions
IdentifyDynamics_Fall2018_forMULTIC.m : facilitates AlternateMinimums_Fall2018_forMULTI.m

COMPARE (2 phenotype)-SINGLE-MODEL ONE-STEP PREDICTION VERSUS TEST DATA (4-WELL)
Script_Revisions_Fall2018_2pMODELvsTestData.m : compares 2-phenotype best fit model to 4-well test data
getPredCellONEMODEL.m : puts model predictions to a cell for later analysis
getPredvsSampSTAT.m : computes p-values to compare model predictions and data (also used for ENSEMBLE MODELING revisions)
getPredULBandONEMODEL.m : computes max/median/min of model predictions for plotting of band
plotBandENSEMBLE.m : plots max/min band of model predictions, dotted median (also used for ENSEMBLE MODELING revisions)
plotDataSampENSEMBLE.m : plots data samples over the time horizon (also used for ENSEMBLE MODELING revisions)

IMPROVED PRESENTATION OF TRAMETINIB AND COMBINATION HYPOTHESES
try_script_hyp_imppres_fall2018rev.m : plots test data samples and prediction curves for Tram, Comb. under
    default constraints, extra constraints. 
    Contains p-values, min/max/median prediction curves (pink band), test data samples (black stars)
 
You can’t perform that action at this time.